(en)A microwave thermometer for accurate and non-invasive measurement of subsurface temperature is described. The microwave radiometer measures the brightness temperature of the specimen. In one example, the system is calibrated using an antenna aimed at the air with another antenna aimed at a subject. This approach dynamically compensates for variable drift of hardware components. An accuracy of 2° F. in 61% of cases may be achieved testing the device on phantoms.

1.ApplicationNumber: US-201414571010-A
1.PublishNumber: US-2015185088-A1
2.Date Publish: 20150702
3.Inventor: RABIEIRAD LALEH
COUSE JOHN MICHAEL

4.Inventor Harmonized: RABIEIRAD LALEH(US)
COUSE JOHN MICHAEL(CA)

5.Country: US
6.Claims:
(en)A microwave thermometer for accurate and non-invasive measurement of subsurface temperature is described. The microwave radiometer measures the brightness temperature of the specimen. In one example, the system is calibrated using an antenna aimed at the air with another antenna aimed at a subject. This approach dynamically compensates for variable drift of hardware components. An accuracy of 2° F. in 61% of cases may be achieved testing the device on phantoms.

7.Description:
(en)CROSS-REFERENCE TO RELATED APPLICATIONS
This application claims priority to U.S. provisional application Ser. No. 61/922,220 filed on Dec. 31, 2013, whose contents are expressly incorporated herein by reference.


FIELD OF THE DISCLOSURE
Aspects of the invention relate generally to animal safety, wellness, and health monitoring. More particularly, some aspects of the invention relate to a viewing and system management system that monitors a pet's health and wellness.
BACKGROUND
Animals are far more stoic than humans and often do not complain or demonstrate pain even while they are making adjustments to accommodate their distress. Through market research, pet owners have made it quite clear that they do not need to be told that their pet is sick, but rather they need to know when their pet is getting sick and what preventative steps they should take in response. For example, if an owner knew her pet was getting sick, she could increase her level of observation (e.g., observe whether the animal is eating, drinking, and/or eliminating normally), increase or decrease certain activities (e.g., walks, etc.), and/or visit a veterinarian.
Similarly, veterinarians have very limited visibility into the health of their animal patients as most clinical encounters between a veterinarian and an animal patient are episodic in nature. As such, during normal checkups veterinarians may not always perform or rely on certain readings such as, e.g., blood pressure, respiration rate/variability, or core temperature (sticking a thermometer in the animal's rectum) because such readings may stress the animal further, may be difficult to perform (blood pressure), and/or are unreliable in a stressful clinical setting (animals may exhibit elevated readings in a veterinarian's office with other animals around—sometimes referred to as “white coat hypertension” or “white coat syndrome”).
Accordingly, some past solutions have attempted to remotely monitor an animal in order to provide an animal owner with data relating to the animal's health status while providing veterinarians further data to assist in diagnosing animal health conditions. However, each of these past solutions suffers drawbacks in that they do not provide a comprehensive view of the animal's health and do not provide an owner and/or a veterinarian with adequate information to determine the animal's health status.
Further, it is desirable to determine an animal's core temperature without using invasive measurement techniques. As an example of a non-invasive technique, microwave thermometry provides a method of detecting internal body temperature patterns and is considered as a way of detecting conditions such as breast cancer, brain temperature and hyperthermia.
The market has focused on infrared (IR) thermometers which provide a signal-to-noise ratio (SNR). However, non-invasive thermometers working in the IR band or other non-invasive thermometers cannot detect energy emitted from inside the body and therefore can only detect the skin surface temperature (which is largely affected by environmental factors and is not an accurate estimation of core body temperature).
The thermal blackbody radiation of the tissue indicates its internal temperature pattern, so characterization of this radiation can lead to the sub-surface temperature of the specimen without the need for any excitation signal that may harm the body. As microwave thermography provides a low-cost, noninvasive, passive, and inherently safe technique for inferring temperature distributions within the body, various approaches have been proposed including:
Temperature dependence of Biological Tissues Complex Permittivity at Microwave Frequencies, D. Faktorová. “Non-invasive estimation of hyperthermia temperatures with ultrasound”, Int. J. Hyperthermia, September 2005; 21(6): 589-600. Non-Invasive Thermometry of the Human Body, Miyakawa, 1996. Non-Invasive Thermometry Using a Chirp Pulse Microwave Tomographic Measurement of Temperature Change in Saline Solution Phantoms of the Human Body, Miyakawa, 1994. Ultra-wideband Microwave-Induced Thermoacoustic Tomography of Human Tissues, Tao et al, 2006. Toward 3D UWB Tomographic Imaging System for Breast Tumor Detection, Romeu, 2010. Thermal Microwave Radiation for Subsurface Absolute Temperature Measurement, Susek 2010. A new correlation radiometer. IEEE Transactions on Antennas and Propagation. Aitken GJM. March 1968. Design of Medical Radiometer Front-end for Improved Performance, Jacobsen et al, 2011. Radiometry System for Noninvasive Thermometry using Temperature Controlled Homogeneous Test Load, Sterzer et. Al, 2009. A 4.4 GHz Microwave Thermometer with Compensation of Reflection Coefficient, Stec 2000. Compensated microwave thermometer for biomedical Measurements, Stec et. al, 1993. An L-Band Microwave Radiometer for Subsurface Temperature Measurement, Tofighi. Feasibility of Noninvasive Measurement of Deep Brain Temperature in Newborn Infants by Multifrequency Microwave Radiometry, Maruyama 2000. Monitoring of deep brain temperature in infants using multi-frequency microwave radiometry and thermal modeling, Maruyama, 2001. http://www.resltd.ru/eng/rtm/ S. Gabriel, R. W. Lau, C. Gabriel, Phys. Med. Biol. 41, 163, 1996. http://niremf.ifac.cnr.it, Italian National Research Council, Institute for Applied Physics, Florence, Italy. Dual-mode planar applicator for simultaneous microwave heating and radiometric sensing. Tofighi, 2012.

A microwave thermometer includes a microwave antenna that electromagnetically couples the emitted power from the material under test with a microwave receiver. The received signal is detected by RF power detection circuitry and analyzed by a processor that measures and characterizes the power and determines the temperature of specimen. As the power emitted from lossy mammalian tissue is very low in the microwave range, a very low noise high gain amplifier front end is helpful.
There are various designs for microwave radiometers. Most common are total power radiometers and Dicke radiometers. A total power radiometer includes a simple amplifier-detector combination that measures the total noise power of the thermal signal. A generalized architecture of a microwave thermometer is shown in FIG. 27 . FIG. 27 shows an antenna 2701 , an amplifier 2702 , a power detector 2703 , and an A/D converter/processing system 2704 as known in the art. The energy is coupled to the system through the antenna. The signal is amplified to a desired level and is passed through other microwave blocks (such as filters). The power detector produces an output analog signal proportional to its input signal power. As the total power inserted to the detector is originated from the thermal noise of the load (tissue), the output signal is proportional to the temperature of the tissue and can ideally be derived from it.
However, as the measured power of the radiometer varies significantly with various factors other than the desired noise source; these radiometers do not provide sufficient accuracy. A significant issue is the gain variation of the amplifiers over time.
An approach to address gain variations is in constant calibration of the system. This approach is implemented in Dicke thermometers. The calibration involves including a temperature reference and routing that reference through the microwave front-end and after the receiving antenna). The two common approaches include a switch-circulator Dicke radiometer as shown in FIG. 28A and a switched amplifier Dicke radiometer as shown in FIG. 28B . These devices include an antenna 2801 , a circulator 2802 , an amplifier 2803 , a power detector 2804 , an A/D converter/processing system 2805 , and references 1 and 2 2807 and 2808 , respectively. The system of FIG. 28A includes switch 2806 and the system of FIG. 28B includes switch 2809 . While the former is intended to remove the effect of amplifier input noise using a circulator, it requires a lossy circulator following the antenna, which decreases the signal strength. A common Dicke switched amplifier radiometer of FIG. 28B also suffers from the switch loss but this can be fixed by inserting an amplifier before the switch as suggested in “Design of Medical Radiometer Front-end for Improved Performance” by Jacobsen et al, 2011.
The issue with this Dicke radiometer is the use of the 50 ohm reference. While the 50 ohm reference is useful with a matching impedance to track the change of thermal noise power with temperature, the 50 ohm termination does not capture the effect of antenna and the offset it generates. As shown in FIG. 29A , the output power from the RF front end does not follow changes in phantoms' lower temperature range when the antenna is followed by a passive component. See also FIG. 34 showing a calibration line for a 50 ohm reference on different days.
Accordingly, there remains a need to provide a microwave radiometer that can help detect a subject's core temperature.
SUMMARY
The following proposes an architecture for microwave radiometers that capture the offset generated by the antenna during calibration. The radiometer described herein may be used on, for instance, a water phantom within a 35-41° C. (95-105° F.) temperature range. An observed error of less than +/−1° C. (+/−2° F.) is helpful.
The microwave radiometer may be part of an overall system for monitoring an animal's health. For instance, one or more aspects of the present disclosure relate to monitoring an animal's core temperature.
In an alternative approach, some aspects relate to monitoring a pet or other animal's health and wellness using two or more sensors in order to provide a pet owner, veterinarian, or other party with content useful in monitoring the pet's overall condition. Also, inferences based on analyses of different signals from different sensors monitoring an animal's vital signs, physiological signs, or environmental factors may also be provided. Some aspects of the disclosure provide a wearable device with embedded sensors whose operation may be governed by various operating modes and/or profiles in addition to the signals from other sensors.
A system and method for monitoring the health of an animal using multiple sensors is described. The wearable device may include one or more sensors whose resultant signal levels may be analyzed in the wearable device or uploaded to a data management server for additional analysis. One or more embodiments include variations of the UWB system to accommodate differences in animals.
The various aspects summarized previously may be embodied in various forms. The following description shows by way of illustration of various combinations and configurations in which the aspects may be practiced. It is understood that the described aspects and/or embodiments are merely examples, and that other aspects and/or embodiments may be utilized and structural and functional modifications may be made, without departing from the scope of the present disclosure.



BRIEF DESCRIPTION OF THE DRAWINGS
A more complete understanding of the present invention and the advantages thereof may be acquired by referring to the following description in consideration of the accompanying drawings, in which like reference numbers indicate like features.
FIG. 1 is a schematic diagram of a wearable device for a pet and its components according to some aspects of the disclosure.
FIG. 2 is a functional block diagram illustrating the various types of information received by the wearable device of FIG. 1 .
FIG. 3 is a schematic diagram of a data management system and the various inputs thereto used in conjunction with the wearable device of FIG. 1 according to some aspects of the disclosure.
FIG. 4 illustrates a collar incorporating the wearable device of FIG. 1 .
FIG. 5 illustrates a cross-sectional view of an animal's neck wearing the collar depicted in FIG. 4 .
FIGS. 6A and 6B illustrate top and side views of an embodiment of the wearable device of FIG. 1 .
FIG. 7 shows a harness incorporating the wearable device of FIG. 1 .
FIG. 8 is a flowchart depicting basic sensor processing according to some aspects of the disclosure.
FIG. 9 is a flowchart depicting processing of more than one sensor according to some aspects of the disclosure.
FIG. 10 is a flowchart depicting a sensor triggering other sensors according to some aspects of the disclosure.
FIG. 11 is a flowchart depicting an illustrative example of how an inference may be formed using readings from different sensors according to some aspects of the disclosure.
FIG. 12 is a flowchart illustrating using readings from sensors from the wearable device and another sensor apart from the wearable device according to some aspects of the disclosure.
FIG. 13 shows a table with sensors and their related information in accordance with one or more aspects of the disclosure.
FIG. 14 is a table with potential master/slave relationships of various sensors identified in FIG. 13 in accordance with one or more embodiments of the disclosure.
FIG. 15 shows an illustrative example of how the activation of the sensors of FIG. 13 may be modified in different operation modes in accordance with one or more aspects of the disclosure.
FIGS. 16A-16G are illustrative examples of various sensors and how their threshold or thresholds, frequency of operation, and granularity may be modified based on different profiles in accordance with one or more aspects of the disclosure.
FIG. 17 shows an example of how various sensor profiles may be modified based on breed information of the animal to which the monitoring devices attached in accordance with one or more aspects of the disclosure.
FIG. 18 shows an embodiment with different operation modes of the wearable device in accordance with one or more aspects of the disclosure.
FIGS. 19A-19B show the order in which operation modes take precedence over profiles based on the embodiment of FIG. 18 in accordance with one or more aspects of the disclosure.
FIG. 20 shows an alternative embodiment with different profiles including profiles replacing the operation modes of the embodiment of FIG. 18 in accordance with one or more aspects of the disclosure.
FIGS. 21A-21B show the combination of different profiles of the embodiment of FIG. 20 with options of profile selection by one or more switches in accordance with one or more aspects of the disclosure.
FIG. 22 shows an illustrative example of how profiles may be selected in the wearable device as well as in the DMS in accordance with one or more aspects of the disclosure.
FIG. 23 shows an illustrative example of relevancy windows of readings of on sensor in relation to other sensors in accordance with one or more aspects of the disclosure.
FIG. 24 shows an example of different techniques for monitoring core temperature including microwave radiometry and microwave thermometry in accordance with one or more aspects of the disclosure.
FIG. 25 shows a display of various conditions of a monitored animal in accordance with aspects of the disclosure.
FIG. 26 shows a specific display relating to one of the monitored conditions of the animal of FIG. 25 in accordance with aspects of the disclosure.
FIG. 27 shows a generalized architecture of a total power microwave radiometer.
FIGS. 28A and 28B show Dicke microwave radiometers with reference values input after the receiving antenna.
FIG. 29A shows measurements on the system of FIG. 28B . FIG. 29B shows an antenna with a radiator and a tuning dielectric inside of the radiating component of the antenna.
FIG. 30 shows an embodiment in which two antennas are used in accordance with embodiments of the invention.
FIG. 31 shows another example of the present invention.
FIGS. 32A and 32B show measured temperature verses actual temperature and error in degrees F. verses the water temperature.
FIG. 33 shows output voltage verses input power of a ZX47-60LN+ power detector.
FIG. 34 shows a calibration line using a 50 ohm reference on different days.
FIG. 35 shows local trending of measurements with the temperature of water.
FIGS. 36A , 36 B, and 36 C show bias (offset) between measurements when using real time calibration.
FIGS. 37A and 37B show how errors increase with lower temperatures using a linear model.
FIG. 38 shows an example of core temperature determinations using a thermocouple of a subject (e.g., dogs) and a comparison to an environmental temperature.



DETAILED DESCRIPTION
In the following description of the various embodiments, reference is made to the accompanying drawings, which form a part hereof, and in which is shown by way of illustration various embodiments in which the invention may be practiced. It is to be understood that other embodiments may be utilized and structural and functional modifications may be made without departing from the scope of the present invention.
General Overview
Aspects of the present disclosure are directed to a device worn by an animal including one or more sensors for monitoring one or more conditions of the animal and/or its environment. In some embodiments, the device may be a collar, harness, or other device placed on an animal by a human (e.g., a pet's owner). The wearable device may include a plurality of components including, e.g., one or more sensors and one or more components used to transmit data as described herein. For example, in some embodiments, the wearable device may include a plurality of contact, semi-contact, and non-contact sensors for obtaining information about the animal, its location, and its environment.
Additional aspects of the present disclosure are directed to analysis of the different sensors. For the purpose of this application, at least two locations at which the sensors are analyzed are described herein. First, the wearable device may analyze the sensor data. Second, a remote, data management system (referred to herein as “DMS”) may process the information from the sensors. In addition, the DMS may process the information from the sensors in conjunction with additional information from sources other than the wearable device including information from ancillary sensors proximate to the wearable device (including stand-alone sensors and sensors attached to other devices, e.g., sensors attached to or part of smartphones). Further, the DMS may receive information from owners who have entered specific information based upon their observations of the animal. In addition, the DMS may receive information from third-parties including RSS feeds regarding ambient weather conditions local to the wearable device as well as data from third-party veterinarians or other service providers. It is appreciated that, in some implementations, the sensors may be analyzed only at one location or analyzed at three or more locations. The health-monitoring system may further use the owner observations of the animal collected through, e.g., companion web/mobile based applications, telephone call center activity/teleprompts, and the like. The owner observations may corroborate measured events (e.g., events measured by wearable device 101 and/or one or more external sensors) to assist in lowering the ongoing rate of false positives and false negatives. For example, in some embodiments, the health-monitoring system may include a mobile weight/size mobile device application which instructs the owner to wave a mobile camera integral to the mobile device across an animal with a pre-identified marker in the field of view. Pre-processed data derived from this action may then be uplifted to the DMS where conclusions can be derived as to the animal's weight and size. Such data is then appended to the animal's record. Other important owner recorded observations may include observable items such as caloric intake, blood in urine, black stools, smelly breath, excessive thirst, white skin patches around the face, recording the disposition of the animal, and the like. For instance, the caloric intake may be monitored by an owner through an application running on a computer or smartphone in which the owner identifies what food and how much is being consumed over what interval.
Further, while described herein as being located remote from the wearable device, the DMS may be located on the owner's smartphone or located on the wearable device based on the respective processing power of smartphone and wearable device. In these alternative embodiments, the “DMS” is identified by its ability to receive content from sources other than the sensors of the wearable device and process that additionally received content for forwarding to the owner and/or veterinarian of the specific animal. These alternative embodiments of the DMS are considered within the scope of the “data management system” unless specifically excluded herein. For instance, if the wearable device is considered the DMS, the wearable device would receive data from its own sensors as well as information from either sensors not located on the wearable device and/or additional content provided by the owner, veterinarian, or third party.
Further, the veterinarian may provide information to the DMS 301 including breed, age, weight, existing medical conditions, suspected medical conditions, appointment compliance and/or scheduling, current and past medications, and the like.
For the purposes of this disclosure, some sensors are described as a specific type of sensor in contrast to a more generic description of other sensors. For instance, while the specification describes the use of a GPS unit providing location information, other location identifying systems are considered equally useable including GLONASS, Beidou, Galileo, and satellite-based navigation systems. Similarly, while the specification describes the use of a GSM transceiver using GSM frequencies, other cellular chipsets may be readily used in place of or in addition to the GSM transceiver. For example, other types of transceivers may include UMTS, CDMA, AMPS, GPRS, CDMA (and its variants), DECT, iDEN, and other cellular technologies.
Also, for the purposes of this disclosure, various sensors and combination of sensors are described as being co-located on the wearable device. However, in various situations, one or more sensors may never be used in a specific version of the wearable device. For instance, GPS-related sensors may not be useful for a version of the wearable device that is only to be used post-surgery in a recovery ward of an animal hospital. Because precise location information is not needed when a veterinarian already knows the location of the animal (or even not useable when in doors), a version of the wearable device with the GPS sensor disabled or not even included may be used. Similarly, other sensors may be disabled in (or never included in) this version of the wearable device where those sensors are not expected to be used. For instance, an RF signal sensor (one that determines if a beacon signal from a base station is above a predetermined threshold) may not be provided in a version of the wearable device where that version of the wearable device is never expected to be used with a base station emitting a beacon signal.
As used in this disclosure, the term “content” is intended to cover both raw data and derived events. For instance, one example of the wearable device as described herein includes a profile/operation mode in which raw data from various sensors are uploaded to a data management on a continuous basis. Another example of the wearable device pre-processes information from various sensors and derives event information from the combination of signals (or lack thereof) from two or more sensors. These derived events are referred to as “device-derived events” as their derived in the wearable device. Similarly, the data management system may also derive events (referred to herein as “DMS-derived events”) from content from the wearable device using only the raw data from the wearable device, the device-derived events, or a combination of both. Further, the DMS may further take into account content from ancillary or third-party sensors to corroborate and/or further enhance the DMS-derived events. For instance, data from ancillary or third-party sensors may include audio files, image files, video files, RFID information, and other types of information. To help correlate the data from ancillary or third-party sensors with data/device-derived events from the wearable device, the data from the ancillary or third-party sensors may include timestamps. These timestamps permit the data management system to use the data from the ancillary or third-party sensors as if that data was part of the data/device-derived events from the wearable device. Further, the information exchanged between the wearable device and the DMS and with third-parties and (as well as with third-party devices) may be performed with industry-standard security, authentication and encryption techniques.
The Wearable Device
FIG. 1 is an overview of wearable device 101 and its components according to some aspects of the disclosure. Wearable device 101 may include several internal components, such as, e.g., ultra-wideband transceiver (UWB) and other sensors described herein at least in FIGS. 13-17 . The sensors are represented in FIG. 1 as classifiable into various sensor types shown as Sensor Types A-F 110 , 111 , 112 / 113 , 114 , and 115 . Although not shown separately in FIG. 1 , the sensors are referred to at times herein as N 1 to Nm, with “m” being the total number of sensors included in wearable device 101 .
As shown in FIG. 1 , wearable device 101 includes a processor 100 (or multiple processors as known in the art) with firmware 102 , an operating system 103 , and applications 104 . The wearable device 101 may also include a storage 105 (e.g., a solid-state memory, Flash memory, hard disk drive, etc.). The wearable device may further include one or more an RF radio, a Wi-Fi radio, a Bluetooth radio, and/or a cellular radio transceiver 107 . The wearable device 101 may further include a local input/output connection (e.g., USB, optical, inductive, Ethernet, Lightening, Fireire, status light or display etc.) 108 , and a battery 109 . For purposes herein, local input/output connection 108 and the radio transceiver(s) 107 are generally considered “outputs” though which information may be communicated to an owner or veterinarian directly (through sound emitter/status light/display 604 of FIG. 6 ), directly to a smartphone (via cellular, Bluetooth, or Wi-Fi or other communication pathways) or though the DMS.
With respect to sensor types A-F, sensor type A 110 refers to the types of sensors that have a sensor input 116 and no other internal components (e.g., simplistic photodiode). Sensor type B 111 refers to a sensor with a sensor input 117 and a processor 118 and storage 119 contained within the sensor type B. Here, sensor type B 111 may store data (at least temporarily) from sensor input 117 and process the data to provide a more meaningful result to processor 100 . For instance, sensor B 111 may be a UWB device for monitoring cardiac activity and the like based on movement of a dielectric material (e.g., a heart muscle or other muscle). Processor 118 may control the operation of the UWB and interpret the results. In addition to monitoring cardiopulmonary activity, the UWB componentry may be used for core temperature determinations and as a communication transceiver for communication with a network as known in the art for short distance, high bandwidth communications.
Further, as shown by dotted line 113 , storage 119 may optionally be associated with storage 105 to the point that processor 118 writes directly and/or reads directly from storage 105 (as being shared between processor 100 and processor 118 ). Raw data from sensor types C 112 and sensor types D 113 are processed by preprocessor 120 before the data being sent to processor 100 . Preprocessor 120 may be any type of known processor that corrects/adjusts/enhances data. For instance, preprocessor 120 may be an analog to digital converter, an analog or digital filter, a level correction circuit, and the like. Sensor type E 114 includes any sensors not specifically identified above that provide results from radar-based signaling (including RF signal strength sensors, Wi-Fi IP address loggers, and the like). Finally, sensor type F 115 includes battery sensors that provide data regarding the charge level and temperature of the battery 109 .
Inputs to the Wearable Device
Processor 100 may be any known processor in the art that performs the general functions of obtaining content from various sources in forwarding it through communication interfaces. The processor 100 may also perform specific functions as described herein. The communication interfaces may include one or more of microwave antennas, an RF antenna, and RFID antenna a cellular radio transceiver, and known hardware interfaces (for instance, USB). For example, processor 100 may direct the transmission on demand of data collected from one or more sensors due to an episodic event or may direct the transmission according to a predetermined schedule or when eventually connected to the DMS where the data is collected in an off-line mode.
With respect to the off-line mode of operation, processor 100 receives raw data from the various sensor types A-F 110 - 115 . Next, depending on the sensor and its current profile and/or operating mode, processor 100 stores content relating to readings from the sensors. In a first example, processor 100 merely stores all raw data from the sensors. In a second example, processor 100 only stores indications that a sensor has provided a reading outside of a normal range. The normal range may be set by the current profile and/or operating mode and may include one or more thresholds for each sensor signal. For instance, an ambient temperature sensor may have upper and lower thresholds of 28° C. and 15° C., respectively. If a reading from the ambient temperature sensor passes one of these thresholds, that event is stored by processor 100 and storage 105 identifying that the ambient temperature exceeded the identified temperature range. In this example, either a binary indication that the temperature range has been exceeded or the actual temperature may be stored in storage 105 . Further, to assist with subsequent analyses by the wearable device 101 or analyses performed by the DMS or third parties, processor 100 also timestamps the indication that the temperature has left the identified temperature range. In a third example, processor 100 may store in storage 105 both the raw data from the sensor leaving and identified range as well as the indication that the identified range has been exceeded. For instance, the indication may be one or more flags stored in storage 105 is associated with the sensor reading, the timestamp, and that the range has been exceeded.
In a further example, processor 100 may operate in a low-power mode when, for example, sensor F (the battery sensors 115 ) identify that the battery is too hot and/or the battery is running low on available power. In this example, sensors that require significant power may be disabled or activated less frequently until the power level has been restored or battery recharged.
Further, processor 100 may accept new software updates and change sensor thresholds, settings, etc., per instructions received from the data management system DMS. The DMS is described below with reference to FIG. 3 . In addition, the owner may modify the thresholds to minimize when he is alerted to various sensor readings from the wearable device. This may be permitted or restricted as minimizing some sensitivity may endanger the animal when the owner should be alerted.
In some embodiments, wearable device 101 may be associated with a base station (not shown). The base station may be capable of charging the battery 115 of the wearable device 101 . Further, the base station may emit a steady beacon signal to wearable device 101 (but optional does not receive communications back from the wearable device 101 ). In some embodiments, the base station may be paired to a plurality of wearable devices 101 (e.g., each worn by each one of a common pet owner's animals). In such embodiments, as known in the art with pairing of wireless devices, each wearable device 101 may be paired to the base station at the time of activation through a unique signal signature. Additionally, in some embodiments, each wearable device 101 may be paired to multiple base stations. One of the benefits of using multiple base stations is that, by comparing the relative strengths of signals from the different they stations, the wearable device 101 may be able to generally identify its location relative to the base stations (e.g., via triangulation).
Optional Location Determination
In some embodiments, wearable device 101 may include a GPS receiver 106 as one example of a sensor. The GPS receiver 106 may turn on once a beacon or other RF signal drops below a threshold level, in response to a sensed episodic event, on demand, or according to a predetermined time schedule. Accordingly, the GPS receiver 106 may not be “always on” (and thus may not, e.g., consume power when GPS readings will not be helpful). By way of an example, if the signal strength of a beacon from base station is high, then the wearable device 101 (and accordingly an animal wearing wearable device 101 ) may be assumed to be located near the base station and thus the GPS coordinates of the animal may not be beneficial to, e.g., the animal's owner. Accordingly, the GPS receiver 106 may remain in an “off” state (e.g., powered down state) until, e.g., processor 100 instructs GPS receiver 106 to turn “on” (e.g., when the signal strength from the base station becomes weak or nonexistent).
The GPS receiver 106 may provide any useful information regarding the status of an animal wearing wearable device 101 including location coordinates of the animal, elevation of the animal, specific satellite acquisition status, and the orientation of satellites. Some or all of this information may be used in sensor logic calculations and reduce GPS thrashing (continuous attempts to acquire signals and thereby draining the battery).
The processor 100 may use location information from the GPS receiver 106 to identify a geo-zone (also refer to as a geo-fence) and determine when the wearable device 101 has left that identified area. For example, when an animal wearing the wearable device 101 is playing off leash in a park, the animal's owner (using, e.g., a personal mobile device), the DMS, or other may prompt the GPS receiver 106 to create an instant geo-zone around the location of the animal wearing wearable device 101 . Accordingly, if the pet wanders too far (e.g., outside of that geo-zone), the owner (via, e.g., a signal sent from cellular radio transceiver 107 to a personal mobile device), the DMS, or other may be notified that the pet has traveled outside of the geo-zone.
In embodiments where wearable device 101 is associated with a base station, processor 100 may determine when, e.g., an RF beacon signal, Wi-Fi signal, Bluetooth signal, or other RF technology signal emitted from the base station drops below a threshold level and, in response, may obtain the location of the device from a GPS receiver 106 and record and/or transmit the location of the wearable device 101 via a cellular radio transceiver 107 , Wi-Fi, Bluetooth, or other technology to a pet owner or veterinarian. Thus, according to one aspect of the disclosure, a location of an animal wearing the wearable device 101 may be easily determined when the animal strays too far from the stationary base station. For non-cellular based radios, if the signal strength falls below a certain threshold or is non-existent, processor 100 may change the transmitting profile of the different modems to make them easier to either locate or connect to various available networks or by a mobile device based application being used as directional finder.
In embodiments which include a base station, the health-monitoring system may further interpret readings coming from base station as described herein. For example, signal strength of a beacon coming from the base station and received at wearable device 101 may be compared to a set of thresholds that have been set by the user or defaults provided/derived by the DMS during setup based on high, medium, and low settings. In some embodiments, during activation of the device and after the owner has set up the base station inside their premises, the user may use a companion application (e.g., smartphone application) and walk around her property holding the wearable device and geo-tag important features of her enclosure/yard/field, etc. At each location the GPS coordinates and beacon signal may be logged and uploaded to the DMS to assist in deriving the optimal safe proximity and geo-zones. The owner may also acquire several other base stations that can be placed in other locations that the animal frequents (e.g. weekend properties, pet sitter, etc.) or placed in several locations of a large and evenly shaped property to create proximity zones of unique shapes.
Wireless Communications
The cellular radio transceiver 107 may be used as one means of transmitting and receiving data at the wearable device 101 . In some embodiments, the cellular radio transceiver 107 may provide presence information on a cellular network and/or signal strength readings to assist in the wearable device's 101 logic calculations to prevent thrashing (continuous attempts to acquire signals). Further, the cellular radio transceiver 107 may provide real-time clock adjustments, and may be used for cellular triangulation by the DMS when GPS signals are not available or are at or below a usable threshold.
Inputs to the Wearable Device
FIG. 2 shows an illustrative example of various inputs usable by the wearable device 101 . FIG. 2 shows RF signal 201 , DMS inputs & triggers 202 , content from mobile companion apps/sensors 203 , GPS-related information 204 , device accessory content 205 , Wi-Fi/Bluetooth/ANT-related information 206 , cellular information 207 , spectrum analyses 208 , sound levels or actual recordings of sound 209 , acceleration 210 , core temperature 211 , RFID (relating to internal/external RFID-radios) 212 , battery temperature/battery strength 213 , cardiopulmonary 214 , ambient humidity 215 , and ambient temperature 216 .
The RF signal 201 may receive signals including adjustable settings and options for, e.g., geo-tagging the boundaries of a pet owner's property, etc. as described above with respect to the beacon signal. In addition or instead of RF antenna 109 , wearable device 101 may include Wi-Fi, Bluetooth, and/or other RF technologies 206 . The Wi-Fi/Bluetooth/ANT-related component 107 is intended to cover local, radio-based communication systems from body-worn to body-wide area networks.
Each may be used in conjunction with a GPS receiver 106 and/or cellular radio transceiver 107 or as a replacement to provide two-way data transmission through paired access points as well as provide presence, proximity, and retrieve time of day information identifying the general location of the wearable device 101 .
Wearable device 101 may further accelerometer providing the acceleration signal 210 . The accelerometer may be used to report levels of specific activities of an animal. For example, readings from the accelerometer may be interpreted as the animal being currently engaged in walking, running, sleeping, drinking, barking, scratching, shaking, etc. The accelerometer may also be used to report the possibility of a high impact event as well as corroborate and/or augment other sensor readings. In some embodiments, the accelerometer may be used to control other sensors (e.g., turn on, turn off, leave a breadcrumb, ignore a reading, etc.). Further, the accelerometer may be used to determine which of a plurality of animals is actually wearing the wearable device 101 . For example, if a pet owner uses a wearable device 101 interchangeably among more than one of her pets, a set of specific attributes pertaining to one of the animals may be created and stored in storage 105 for each pet. Some of the stored attributes may be accelerometer data, such as a particular animal's gait, and other attributes such as bark sound signatures. These stored attributes may then be used to determine which pet is wearing a wearable device 101 by comparing currently sensed attributes to stored attributes.
Another sensor usable with the wearable device 101 is a light meter. The light meter provides the spectrum analyses 208 input of FIG. 2 . In a simplistic example, the light meter may be tied solely to presence or absence of a threshold of visible light. In a more sophisticated example, the light meter may be frequency-specific in its readings such that it can separately detect levels of infrared light, visible light, and ultraviolet light. Both of these examples of light meters of varying sophistication are known in the art. In this environment, the processor 100 uses signals from the light meter (or light meters) to determine if the wearable device 100 is located inside or outside. For instance, while a visible light level of a given intensity may indicate that the wearable device 100 is located under a bright light source (e.g., in a sunny area), processor 100 may compare the current infrared and/or ultraviolet light levels against the visible light levels. Accordingly, if the visible light level is high and the infrared and/or ultraviolet light levels are also high, then processor 100 determines that there is a likelihood that wearable device 101 is located outside in the sun. Alternatively, if the visible light level is high while the infrared and/or ultraviolet light levels are low, then processor 100 determines that there is a likelihood that wearable device 101 is located indoors (albeit in a sunny spot).
Further, the light meter may also be used to interpret light levels in determining a current state of an animal to confirm or corroborate a current state of an animal. For example, in some embodiments extremely bright light incidences may be indicative of the animal wearing wearable device 101 being caught in a car's headlights, or being around gunfire, explosions, etc. as based on the sudden change in received light levels 208 . Identification of being caught in a car's headlights may be based on a sudden spike in ambient light at night while the accelerometer indicates minimal movement before and after the spike in visible light. Further, a location determination (for instance, from a GPS receiver) may be used in place of or in addition to the accelerometer signal as augmenting the determination of whether the animal has been illuminated by oncoming headlights. Similar spikes in audio signals occurring within a short time of visible light spikes may be interpreted as being around the gunfire, explosions, etc.
More advanced uses of spectrum analysis include the ability to detect trace chemical signatures present in the animal's environment, emanating from their skin/fur, orifices, and/or present in their breath. For example, readings could indicate dangerous environmental conditions (e.g. high readings of chorine), skin related issues (e.g. yeast), and internal related conditions (e.g. ketones in the animal's breath that may be exhibited before other symptoms are evident). Further, the spectrum analysis sensor may also be sniffing for chemical signatures. Combining the detection of sulfur with light and sound spikes helps corroborate the determination that the animal has recently been located near gunshots or other explosions.
An ambient temperature sensor providing the ambient temperature 216 may also be provided as another example of a sensor. The ambient temperature sensor may be used to determine a location of an animal wearing wearable device 101 (e.g., indoor versus outdoor). In some embodiments, the processor 100 tracks ambient temperature 216 over time and determines a current rate of change. If that current rate of change is greater than a predetermined rate as existing for a period of time, processor 100 identifies the rate of change is a prediction that the animal wearing wearable device 101 will be overheating or freezing in the near future. Further, in some embodiments an ambient temperature sensor may be used to corroborate or control other sensors.
The wearable device 101 may also include a humidity sensor providing the ambient humidity input 215 . In some embodiments, the humidity sensor may be used to adjust sensed temperatures to wet bulb settings. These wet bulb settings may be important in calculating animal heat loss/gain and may be used in roughly identifying a location of the animal (e.g., inside or outside). Further, the excessive humidity or dryness identified as signal 215 from the humidity sensor may be combined with a temperature reading to determine the heat index or wind chill.
Further, a microphone or peak noise detector sensor may provide sound input 209 . The microphone/peak noise sensor may be used to, e.g., measure specific sound events (barking, etc.) and may be used to corroborate other sensor readings. For example, in embodiments where a light meter indicates, e.g., an animal wearing wearable device 101 may be caught in a vehicle's headlights; a microphone sensing a load noise may be interpreted as, e.g., an impact event (getting hit by the vehicle). A specific method of determining an impact event is described herein.
Another example of a sensor may be an internal battery strength and/or battery temperature sensor 213 providing information regarding the strength and/or temperature of the battery. The internal battery strength and/or temperature sensor may be used to either modulate certain other sensing activities and/or as an input source to other sensing activities. For example, in response to sensing the internal battery is running low, GPS acquisition duty cycles and/or cellular transmissions may be reduced to conserve power to extend the operation of the wearable device 101 .
A core temperature sensor providing core temperature 211 may be provided as another example of sensor. The core temperature sensor may be used to non-invasively measure the core temperature of an animal, and thus provide data relating both to a real-time core temperature of an animal and an animal's change in core temperature over time.
The wearable device may also include one or more antennas as tied to one or more of the internal radios/sensors. One of the internal components attached to the antennas may be a UWB device. As known in the art, UWB device is used to monitor various conditions (e.g., used in fetal monitoring, cardiopulmonary monitoring, and the like). Here, the UWB device may be used to monitor a variety of different conditions. For example, in some embodiments, the UWB device may be used to transmit and receive UWB signals to non-invasively monitor operations of an animal's heart. Signals from that monitoring operation are then processed by processor 100 to determine if an episodic event has occurred (e.g., an abnormally high heart rate), if a more complex event has occurred (e.g., heat exhaustion after excessive running) and if the cardiopulmonary system of the animal is trending toward an undesirable condition (e.g., an increasing average heart rate). Here, in addition to an average heart rate, a statistical deviation may also be provided. In this regard, statistical deviations may accompany other average rates as forwarded to veterinarians and possibly owners.
Specifically, the UWB device may be used to measure stroke volume and a relative change in blood pressure of an animal wearing wearable device 101 . For purposes herein, stroke volume readings from the UWB are useful in addition to vital sign readings. In other embodiments, the UWB device may be used to determine if the wearable device is actually on the animal. In some embodiments, a profile (e.g., stored characteristics) of an animal may be available for more than one animal which wears the wearable device 101 . In such embodiments, the UWB device may be used to determine to which animal the wearable device 101 is currently attached. For example, readings at the UWB device may be compared to stored cardiopulmonary profiles to determine which of a plurality of animals is currently wearing the wearable device 101 . Further, the UWB device may be used to interpret changes in the neck tissue as indicative of an animal eating, drinking, and/or vomiting. Further, the UWB device may be used to interpret signals in the abdomen area to investigate the possibility of obstructions in the digestive track.
Any other desirable sensor may be provided as a component of wearable device 101 in order to measure one or more attribute of an animal and/or its environment. Those skilled in the art, given the benefit of this disclosure, will recognize numerous other sensors which may be incorporated into wearable device 101 without departing from the scope of this disclosure. Further, the components and/or sensors contained within wearable device 101 may share some common circuitry such as power supply, power conditioners, low pass filters, antennas, etc., as well as share sensing data with each other to derive more meaning from combined data sources.
According to some aspects of the disclosure, the wearable device 101 (and associated base station(s), if any) and the DMS may form part of a health-monitoring system used to collect data about and/or monitor specific health attributes of one or more animals. Further, in some embodiments, one of more of sensors may have the capability of activating, deactivating, controlling, rejecting, accepting, or throttling another sensor's activities as described herein. In addition, the health-monitoring system may include both passive and active sensors and multiple antennas that generate and receive a wide variety of electromechanical energy whereas the normal output of one or more components may enhance the capability of another component in a derived fashion.
The health-monitoring system according to some aspects of the disclosure may further include external sensors (e.g., sensors external to the wearable device 101 ) which interact with or otherwise supplement the sensors of the wearable device 101 . In some embodiments, these external sensors may include detachable analog/digital items such as a stethoscope, ultrasound sensor, infrared temperature sensor, pulse oximeter, blood pressure monitoring tool, glucose meter, blood analyzer, breath analyzer, urine analyzer, brain scanner (all which may include additional application software and/or be controlled by the device software), and filters/attachments to enhance/collaborate the existing set of sensors and readings. The individual operations of these separable sensors are known in the art. Here, wearable device 101 provides a platform to which these additional sensors may be connected and their data or analyzed content being stored in storage 105 for relaying to an owner or DMS (or even third parties) as described herein.
In some embodiments, these external sensors may be integrally provided with or associated with other well-known devices. For example, the health-monitoring system may collect data from a camera (with or without lens/filter attachments), microphone, speaker, GPS, and other items that may be plugged into or utilized by the wearable device 101 and/or the health-monitoring system. In some embodiments, these sensors may be part of a personal mobile device (e.g., a smartphone or the like). Each of these external sensors and/or mobile browser applications/installed applications may act independently, in conjunction with the wearable device 101 , may be triggered by the wearable device 101 , or may be triggered by the DMS on a demand, episodic, or a scheduled basis to provide additional and/or collaborative sensing information that will provide important episodic, derived, or trending information to support the animals safety, wellbeing and health. In addition, all of the above described activities may be triggered by a mobile device and a companion applications and attachments/accessories to provide time stamped correlation of sensor data as described herein.
Further examples of external sensors used in conjunction with the described health-monitoring system may include RFID proximity sensors that communicate with RFID proximity tags and provide RFID content 212 . For example, RFID proximity tags may be placed at an animal's bed, at its food bowl, at its water bowl, outside a door frame, outside a gate post, near garbage cans, etc. Thus, when an animal wearing a wearable device 101 is near any of the above items, the wearable device (receiving a signal via the RFID sensor) may interpret that the animal is sleeping, eating, drinking, outside, out of the yard, getting into garbage, etc.
The health-monitoring system may further use owner observations of an animal collected through, e.g., companion web/mobile based applications, telephone call center activity/teleprompts, and the like. The owner observations may corroborate measured events (e.g., events measured by wearable device 101 and/or one or more external sensors) to assist in lowering the ongoing rate of false positives and false negatives. For example, in some embodiments, the health-monitoring system may include a mobile weight/size mobile device application which instructs the owner to wave a mobile camera integral to the mobile device across an animal with a pre-identified marker in the field of view. Pre-processed data derived from this action may then be uploaded to the DMS where conclusions can be derived as to the animal's weight and size. Such data is then appended to the animal's record. Other important owner recorded observations may include observable items such as caloric intake, blood in urine, black stools, smelly breath, excessive thirst, white skin patches around the face, recording the disposition of the animal, and the like. For instance, the caloric intake may be monitored by an owner through an application running on a computer or smartphone in which the owner identifies what food and how much is being consumed over what interval.
Further, the health-monitoring system may include sensors placed internally within an animal (for instance, invasive but unobtrusive sensors). For example, microchips or the like embedded within an animal may provide data relating to, e.g., blood oximetry, glucose monitoring, ECG, EEG, etc.
Data Management System
FIG. 3 shows an example of a data management system 301 receiving inputs from a variety of sources. Those inputs may be specific to an individual animal or generally relate to related animals (related by one or more characteristics including breed, age, health condition, and the like). FIG. 3 shows data management system 301 receiving RSS feeds 302 , Internet search content 303 , social form content 304 , content from chats with veterinarians, symptom lookups and the like 305 , cellular network-related information 306 , Wi-Fi/Bluetooth/ANT-related information 307 , wearable device 101 -based sensors and accessories 308 , third-party electronic services 309 , veterinarian observations 310 , content from companion mobile apps/sensors 311 , owner observations 312 , and third-party home tele-health sensors 313 .
DMS 301 is a data receiving and processing system that receives data and/or wearable device-derived events from the wearable device 101 and analyzes that content directly, or in conjunction with older data or past analyses of older data from the wearable device, or in conjunction with data from other sources, or any combination thereof. The DMS 301 includes one or more processors, storage, operation software, input/output pathways, and the like as similar to that of the processor 100 and storage 105 of wearable device 101 shown in FIG. 1 . Further, the DMS may be a cloud-based computing platform in which communications via the Internet are received in the DMS at a server or other hardware device and processed in accordance with computer-executable instructions and workflows. In this example, the DMS may have industry standard Internet connections, routers, servers, that connect DMS 301 to the various content sources 302 - 313 . Alerts as sent to an owner compared to a veterinarian may be different. Further, even if the sensors are operating as tied to a specific profile, the DMS may continue to separate and forward alerts based on predefined settings at the DMS.
In some embodiments of the disclosure, the health-monitoring system may further collect data using external rich site summary (RSS) feeds 302 . For example, the system may receive data about the weather, environment, daily pet health tips, published research data, etc., via the RSS feed 302 . According to some aspects, this received data may be used to corroborate, supplement, and enhance data collected from the wearable device 101 , other external sources, and the like as discussed herein.
Some embodiments of the health-monitoring system may further receive data from, e.g., non-invasive home telematics solutions 313 . For example, the system may receive data from smart mats, smart motion/IF detectors, and other devices prevalent in the marketplace. Pets and animals inside a home may thus trigger these devices and thus record sensor artifacts such as presence, weight, physiological signs, and vital signs. These recordings (which may normally be discarded by the human home monitoring systems) may provide valuable data collection/corroboration points for the system, for example in the DMS (as described herein). Several techniques may be employed to upload this data to the DMS (e.g. companion mobile device application, user-entered readings, Bluetooth, Wi-Fi, other RF technologies, etc.).
When used as part of a health-monitoring system in FIG. 2 and as described herein, the wearable device 101 may be the prime source of sensor collected data (through, e.g., sensors and others described above). All sensors and their inputs may be available to be intelligently combined through data fusion to create meaningful standalone alerts and as an input into the DMS to develop and extract even more meaning from the data.
In some embodiments, the health-monitoring system as described herein may include a DMS 301 remote to the wearable sensor 101 as schematically depicted in FIG. 3 . In some embodiments, DMS 301 may receive information from wearable device 101 and/or other sensors. Further, DMS 301 may transmit information to, e.g., a pet owner (via, e.g., a computer, smartphone, tablet, land line, display of wearable device 101 , status light/display/sound indicator 604 of FIGS. 6A and 6B , etc.) and/or a veterinarian (via, e.g., a web-based dashboard, facsimile, land line, mobile alerts, etc.). In some embodiments, DMS 301 may transmit data according to predefined criteria. For example, according to some aspects, DMS 301 may transmit information periodically on a scheduled basis. In other embodiments, DMS 301 may transmit information when that information exceeds a threshold value. In still other embodiments, DMS 301 may transmit data on-demand (e.g., requested by a pet owner, veterinarian, or the like).
In some embodiments, DMS 301 may be the data repository of all inputs regardless of the source to derive meaningful/actionable information related to the animal's safety, wellness, and health for owners and veterinarians. In some situations, information specific to the animal wearing the device 101 (e.g., the third-party information service data 309 and the third-party veterinary chat service data 311 ) may be forwarded from the DMS 301 to the third-party prior to receiving data ( 307 , 311 ) from the third parties to assist with the third-parties' analysis. The DMS may analyze received data and determine the meaning of the data as DMS-derived events. Next, based on those events, the DMS may obtain recommendations on file from a storage tied to those derived events, compile those recommendations, and provide the compiled recommendations to the owner and/or veterinarian as actionable information. For instance, if the meaningful information is that the animal has gained 5 lbs. in the past week and has exhibited a lower than normal activity rate, the DMS 301 may look up recommendations on file from a storage tied to weight gain and the amount of weight gain and the identified recommendation or recommendations. Next, the results are compiled and forwarded to the owner/veterinarian as actionable information.
In general, the following lists typical inferences that may be reported to owners: the animal is outside of designated safe zones; there is a potential situation where the animal may be overheating or freezing; the animal may have been in an accident (high impact event of various levels of severity); the animal's activity level has been decreasing even after applied filters for owner and pet lifestyle profiles; the animal is limping (based on a change in gait); the animal appears to be in potentially dangerous environment based on extreme noise and light indicators; the animal is very listless during sleep (as an indication of pain, digestive issues, respiration issues, or past physiological trauma); the animal's heart rate variability is abnormal; the animal's respiration rate and quality is abnormal; the animal appears to be in distress/pain (yelps when there is large gross movement); and the wearable device is not on the animal that it was initially assigned to by means of examining its gate profile versus the one on file or other vital sign indicators that are part of their electronic profile.
Typical suggested actions may include to: increase the owner's personal observations of the animal to confirm or dismiss specific developing items of concern; increase/decrease thresholds for items in the animal's sensor profile so they more closely align with the owner's and the specific pet's daily life patterns, age, breed, size, and know medical conditions; increase/decrease the animal's activity; monitor the animal's diet (record caloric intake); remove the animal from a potential developing overheating/freezing situation; monitor the animal for specific coughing sounds; refer the owner to specific related articles/links/videos etc.; consult an optional online “ask-a-vet” services; and to see their veterinarian as soon as possible based on a life-threatening situation.
The following are illustrative examples of triggers that result in reporting issues to the owner: an episodic issue based on a sensor or a group of sensors confirming an event comparing readings to preset thresholds; a time-based analysis (a.k.a a longitudinally-based) on analysis at the device 101 level or the DMS 301 level based on trending positive or negative readings for a particular suspected condition; on the demand of the owner or the veterinarian; periodically to provide a snapshot of the condition of the animal based on the owner or veterinarian's safety, wellness and health goals.
The veterinarian may receive a fewer number of inferences/suggestions and more empirical data based on wellness issues and vital signs that could lead to serious health issues, the monitoring of specific known health conditions, and the monitoring of the effectiveness of prescribed therapies. The veterinarian may receive vital signs and other physiological information that suggests the animal is trending positively or negatively. Items that may act as triggers for the veterinarian to be sent information include an episodic vital sign(s) reading or physiological reading has passed its threshold or a derived vital sign(s) or physiological sign or signs as trended over time have passed thresholds set by the veterinarian.
Also, the veterinarian may be interested in the following current possible vital, environmental, or physiological signs: core temperature; ambient temperature & humidity; and core temperature. The veterinarian may be interested in the following pulmonary information: detected lung motion & measured respiratory rate and rhythm; measured respiration and exhalation times (ti/te); detected asymmetrical respiration (inflammation, obstructions, asphyxiation); measured chest compression rate, depth, and chest recoil; and measured and ongoing monitoring of chronic bronchitis. The veterinarian may be interested in the following cardiac information: detected cardiac motion & measured cardiac rate and rhythm; measured changes in cardiac stroke volume and cardiac output; a comparison of blood pressure to a threshold; signs of developing congestive heart failure; signs of bradycardia and tachycardia; signs of hemo/pneumothorax. Further, the veterinarian may be interested in the following other information: signs of a seizure; uterine contraction rate and intensity; identification of possible sleep problems such as sleep apnea; signs of a foreign body in the animal; long-term sensor data; average and statistical deviation of cardiac activity, respiration activity, and core temperature; activity level; estimated weight; estimated hydration levels; and average daytime/nighttime ambient temperatures. The following are sample inferences that may be derived by the DMS 301 and identified to the owner or veterinarian for diagnosis: heartworm; vomiting & diarrhea; obesity; infectious diseases; kennel cough & other developing respiratory conditions; lower urinary tract infection; dental disease; skin allergies; damaged bones & soft tissue; cancer (for instance, by ketone level changes in the animal's breath); developing heart conditions; distress/pain; and cognitive dysfunction. The following are sample symptoms/inferences made from a combination of sensor data and veterinarian-supplied data: impact of specific prescribed therapies; recovery status of an animal who has just undergone surgery; and trending of vital signs against a base line determined by the veterinarian.
In such capacities, the DMS 301 may be receiving raw data, pre-processed data at the wearable device 101 level. For example, the accelerometer {x,y,z} g values may be averaged over a fixed window (for instance, a one second window), a deviation of magnitude computed, and a high, medium, or low activity designation may be assigned based on the activity of the animal. Sound files from a separate device, RSS feeds, and other unlike data types need to be catalogued, time stamped, sorted and prepared for analysis. Because the DMS receives these divergent types of data, the DMS 301 may perform these correlations. For instance, the DMS 301 may receive high ambient temperature readings from the wearable device 101 and compare it against expected local temperatures (obtained by RSS feed 302 or Internet search 303 ) for the current or last identified location of the wearable device 101 . If the ambient temperature is high (for instance, over 45° C.) while the predicted high temperature for the location is only 20° C.), then the DMS 301 may derive that the animal is locked inside a car with its windows shut. Based on this derived event, the DMS may attempt to alert the owner as alert 314 . The alert 314 may be in the form of one or more emails, SMS or other text messaging systems, social messaging systems (like Twitter and Facebook, etc.) or by calling the owner directly. It is appreciated that the frequency and thresholds for alerts may be fixed or may be configurable by the user.
DMS 301 may also include information about past events, current events, or predictions of possible future events. DMS 301 may also act as the communications hub between the wearable device 101 and third party services, the vet, and/or a pet owner through various communications channels and devices. For example, in some embodiments a pet owner may use her personal mobile device as an input device to record her own observations through free form text or drop down menus (effectively becoming one sensor of the sensory platform) and thus DMS 301 receives these inputs from the owner via the personal mobile device. Each data element stored in the DMS 301 may be meta-tagged so that each stands alone without having to go back to, e.g., an owner/pet profile. Such meta-tags may include a time stamp, geographical data, breed, age, etc., that may facilitate large scale anonymous data analysis.
Neck Placement of Wearable Device 101
FIG. 4 illustrates a collar 402 including wearable device 101 according to one aspect of the disclosure. As depicted in FIG. 4 , collar 402 may include wearable device 101 such that the wearable device 101 is positioned near the neck of animal 401 . Accordingly, in such an embodiment, sensors receive data near the neck of animal 401 at sensing location 402 . Further, wearable device 101 receives and transmits data at transceiving location 404 .
FIG. 5 illustrates a cross-sectional view of animal's neck wearing collar 402 including wearable device 101 . As depicted, collar 402 may include a clasp 505 that, when clasped, positions wearable device 101 adjacent to fur 501 on the lower side of animal's neck. FIG. 5 depicts approximate locations of the structures within the animal's neck. Specifically, FIG. 5 shows carotid arteries 503 , jugular veins 504 , esophagus 509 , trachea 511 , and spinal column 510 in relation to the wearable device 101 . In such a configuration, antennas of the cardiopulmonary (e.g., UWB device) and other inward-looking components (e.g., ECG and ultrasound probes) contained in wearable device 101 are placed on the inside of collar 402 while processor 100 , other sensors, and other components (e.g., RF antennas 109 , RFID antennas 111 , etc.) are located on the other side of collar 402 (for instance, at location 507 ). Further, the outward looking antennas may be located at any of locations A-I to help minimize interference with the inward-looking antennas. Alternatively, sensors located at locations A-I may have improved readings by separating them from interference with contact with the animal. For instance, if the ambient temperature sensor was placed at location A, there is a potential for errant readings when the animal is laying on its chest and wearable device 101 is resting on the animal's paw. Locating the ambient temperature sensor at an alternative location, for instance, D-I, may improve the reading from the sensor as it would be spaced from the animal's paw when the animal is laying in this position. Further, in an alternative example, various sensors may be replicated around the collar 402 and their readings averaged or the highest and lowest readings dropped to reduce the influence of aberrant readings.
As shown in FIG. 5 , wearable device 101 is able to receive and transmit information on the outside of collar 402 , while keeping inward-looking antennas near animal's skin on the inside of collar 402 such that accurate readings from, e.g., the animal's carotid arteries 503 and/or esophagus 509 may be obtained. Alternatively, readings may be obtained from jugular veins 504 instead of or in conjunction with carotid arteries 503 . Other tissue movement may also be of interest including muscle movement surrounding the trachea (as the trachea's cartilage may not be reflective of some dielectric signals and not detectable directly).
The configuration of wearable device 101 according to some embodiments of this disclosure may be more readily understood with reference to FIGS. 6A and 6B . FIG. 6A illustrates a top view and FIG. 6B illustrates a side view of an embodiment of wearable device 101 . In the embodiment of FIGS. 6A and 6B , wearable device 101 may include two portions: an inside portion 601 and an outside portion 603 . Inside portion 601 may include the inward-looking antennas such as the UWB antennas, microwave antennas, or ultrasound antennas. For instance, the antennas may be located at locations 605 and 606 . Outside portion 603 may include other components such as processor 100 and the other components of FIG. 1 including outward-looking antennas. In one example, the inward-looking antennas of portion 601 may be shielded from the outward-looking antennas of portion 603 by a metal or metallized layer or other known antenna isolation material to minimize interference between the different sets of antennas. Further, status information including on/off status may be provided to the owner via status light 604 . Status light 604 may be a simple LED or may include a display screen and touch interface configured to display content to an owner as opposed to (or in addition to) sending the information to the DMS to then be forwarded to the owner's smartphone. In addition, 604 may be a sound generator that responds to setting changes.
When wearable device 101 is placed on an animal, such as shown in FIG. 5 , the inward-looking antennas will be located near the animal 401 (e.g., inside of collar 402 ) and thus provide accurate sensing, while other components, including some components used to transmit and receive data, may be placed away from animal 401 (e.g., outside of collar 402 ) such that transceiving capabilities of the outward-looking antennas are not degraded by the operation of the other antennas.
Further, metal or metallized probes 610 and 611 may be used to establish probe-to-skin contact for sensors that may be improved with direct skin contact. These types of sensors may include skin temperature sensors, heart rate sensors, and ECG sensors. With respect to temperature sensors, these probes may be attached to one or more heat-sensing components (or may include those heat-sensing components. The heat sensing components may include thermistors, thermocouples, and the like and combinations thereof.
Chest Placement of Wearable Device 101
In other embodiments, wearable device 101 may not be worn around a neck of an animal 401 , but rather may be worn at any suitable location for receiving information by the sensors. For example, and as illustrated in FIG. 7 , wearable device may be provided as part of a harness 701 worn around animal chest. In such an embodiment, sensing location 703 and transceiving location 704 will be near animal's chest rather than near animal's neck (as depicted in FIG. 4 ). Regardless of the particular location of wearable device 101 (at the neck location or chest location, batteries 115 and other detachable components may be removable and replaceable by a pet owner 705 .
Operation of Sensors
FIGS. 8-12 and 22 relate to flowcharts showing processing of the wearable device 101 and/or DMS 301 . These flowcharts are used to explain various aspects of analyzing signals from one or more sensors. It is appreciated that other types of analyses based on the sensor information are possible in place of threshold comparison. Other known techniques include Bayesian inference analysis, neural networks, regression analysis, and the like and their use to analyze the signal inputs are encompassed within the scope of this disclosure.
Turning now to FIG. 8 , a flowchart representing basic sensor processing (e.g., processing of one or more internal sensors, external sensors, internal sensors, and/or other sensors) is depicted. A sensor processed as shown in FIG. 8 may be one that is either on all of the time, interrupt driven, or triggered on demand. At step 801 , sensor data is received from sensor n. Again, this sensor data may be continuously received (e.g., always on), may be triggered by another sensor's reading (e.g., interrupt driven), or may be received in response to a pet owner, veterinarian, or the like requesting sensor data (e.g. on demand). At step 803 , the received sensor data is compared to a threshold value. At step 803 , the relationship of the compared data to the threshold value may be such that nothing of interest is happening. In such a situation, the data may be ignored as indicated by step 809 , and the method will return step 801 to receive additional data. However, if the compared data exceeds the threshold, this occurrence is written to storage in step 805 . Optionally or in addition to step 805 , an alert may be provided to a pet owner or sent to the DMS as shown in step 807 . The alert may be local (e.g., an audible alarm on the wearable device 101 ) and/or may be remote (e.g., on a pet owner's personal mobile device, within a veterinary dashboard, etc.). In a further modification, the fact that the signal from sensor n did not exceed the threshold may also be stored as shown in broken lines from the NO output of determination step 803 to the ignore step 809 as a positive indication that the reading was within the threshold. Further, the series of store ratings provide a breadcrumb data set of incremental changes that may be usable by the DMS.
FIG. 9 depicts an embodiment where readings from multiple sensors {n 1 , n 2 , and n 3 } may be used to determine a status of an animal. Again, each of the sensors in the diagram may be constantly on, interrupt drive, or triggered on demand. At steps 901 , 903 , and 905 , data is collected from each sensor n 1 through n 3 . As discussed, the sensors may be located in wearable device 101 and/or external devices (e.g., smartphone, RSS feed, etc.). Any one of sensors n 1 , n 2 , and n 3 may individually trigger an alert condition in step 906 , and written to storage in step 907 and (optionally) the alert provided to the owner or DMS in step 909 . Otherwise, the determination is ignored in step 908 . Similar to the process of FIG. 8 , data may be breadcrumbed despite the sensor readings not exceeding a threshold as shown in the broken lines from step 906 to step 907 and then back to step 908 .
Alternatively, step 906 may require a consensus of all three readings a weighted basis is needed to either confirm an alert condition or ignore the sensed the data. For example, at step 907 , in response to one or more of sensors n 1 , n 2 , and/or n 3 triggering an alert condition at steps 901 , 903 , and/or 905 , respectively, a combination of the sensed data from each sensor is compared to one or more thresholds to determine if, e.g., an alert condition is present. Further, at step 907 the sensed readings may be compared to past readings that are either stored locally (e.g., within wearable device 101 ) or stored, e.g., in the DMS 301 . Thus, using the sensed data from multiple sensors (in the depicted embodiment, n 1 through n 3 ), inferences regarding animal and pet safety, wellness, and health may be formed at step 907 based on analysis of the sensor's readings and/or, e.g., breadcrumbs (time-stamped recordings). If the combination of the sensor data triggers an alert (e.g., if the combination of data confirms an alert condition), the alert may be returned at step 909 (to, e.g., a pet owner and/or veterinarian, etc.). However, if the combination of sensor data does not trigger an alert after being compared to one or more thresholds, the data is ignored at step 908 and the method returns to steps 901 / 903 / 905 to receive further data. In any event (e.g., alert or ignore) the readings and results may be written to local storage at step 907 for subsequent upload to the DMS 301 .
The analysis of the sensor data at step 803 or the multiple sensor data at step 907 may be performed in any suitable location within the system. In some embodiments the analysis may be performed in the wearable device 101 . In such embodiments, wearable device 101 may perform episodic data analysis (e.g., independent intelligent decisions) as well as longitudinal data analysis. For the latter, the wearable device may monitor a number of recorded breadcrumbs of various events over time. For example, the wearable device 101 may monitor the animal's temperature over time in order to monitor the animal's condition in compliance with FAA regulations on pets stored in cargo holds. In other embodiments, the wearable device 101 may monitor the animal's barking over time to ensure the animal 401 is complying with local by-laws or to interpret continued barking as a potential stress indicator.
In other embodiments, the analysis of the sensor data may be performed in DMS 301 . Again, DMS 301 may perform both episodic data analysis as well as longitudinal data analysis. For the latter, DMS 301 may look at individual events, combined events, and derived events (e.g., calorie intake versus activity levels). By looking at such events in the DMS 301 , patterns of animal's 301 health and wellness may be determined. For example, the DMS 301 may determine patterns of improvement (or lack thereof) of an animal following a drug or therapy treatment of animal 401 after it has left the veterinarian. Further, the wearable device 101 data may be combined with sensors from other sources (e.g., RSS feeds 302 , owner observations 312 , etc.) in performing the analysis. For example, an RSS feed 302 including the number of degree days may be compared to a number of high temperature alerts at a wearable device 101 to determine if, e.g., animal 401 is overheated or if, rather, it is just an abnormally warm month. As another example, owner's observations 312 (e.g., observations of staggering after exertion, unusual fatigue, abnormal coughing, pale gums, etc.) may lead the DMS 301 to modify the profile or operation mode of the wearable device to employ profiles with finer granularity and sensing more often and with more sensitive thresholds for cardiopulmonary algorithms at the wearable device 101 level.
As presented in FIGS. 8 and 9 , an analysis of an animal's health and wellness may be performed by analyzing data from an individual sensor (e.g., FIG. 8 ) or from the combination of two or more sensors reading at the same time (e.g., FIG. 9 ). In other embodiments, analysis of an animal's health and wellness may be performed by one or more sensors triggering one or more additional sensors in order to corroborate the data of the first sensor. This may be more readily understood with reference to FIG. 10 . As shown in FIG. 10 , data is received from one sensor (in the depicted embodiment, n 1 ) at step 1001 . This data is compared to one or more thresholds at step 1003 as described with respect to FIGS. 8 and 9 . If the sensor reading does not exceed a threshold (e.g., is not interesting) then the data is ignored at step 1007 and the method returns to step 1001 to obtain additional data. Alternatively, the data may always be stored/written locally at step 1005 for later upload to DMS 301 .
If the data from sensor n 1 obtained at step 1001 does exceed one or more thresholds at step 1003 , then signals from additional sensors may be checked to confirm or corroborate the received data from step 1001 . That is, in some embodiments, one or more sensors (in the depicted embodiment, n 1 ) may act as a “master” sensor after it has sensed a threshold level, and then subsequently control additional “slave” sensors. Here, steps 1001 - 1009 are related to the operation of the master sensor n 1 , collectively identified by the dashed box 1000 M. Similarly, steps 1010 - 1014 are related to the operation of the slave sensors n 2 and n 3 , collectively identified by the dashed box 1000 S. In the depicted embodiment, once data collected at step 1001 exceeds a threshold at step 1005 , additional slave sensors are triggered to collect data at step 1010 (n 2 ) and step 1011 (n 3 ) or their previously collected data checked. At step 1012 , analysis of the received data (e.g., data received at steps 1001 , 1010 , and/or 1011 ) may be performed, and an inference may be made regarding animal's health and wellness. Further, the data received from each sensor (n 1 , n 2 , and n 3 ) may optionally be weighted or otherwise adjusted to determine an inference regarding an animal's health and/or wellness as described herein. If, at step 1012 , the combined data does not exceed a threshold level (e.g., the further data collected at steps 1010 and/or 1011 does not confirm and/or rather negates an inference made at step 1003 ), then the data may be ignored at step 1007 and the method thus returns to step 1001 to collect new data and thus continually monitor animal 401 . However, if the data collected at steps 1010 and/or 1011 confirms or supplements the inference made from the data collected at step 1001 , then this determination is recorded in step 1013 by writing this determination into storage 105 . Further, an alert may be returned to the animal's owner and/or a veterinarian at step 1014 . Again, regardless of the inference made (e.g., ignore versus alert) the data may be written/stored locally at step 1013 for future upload to the DMS 301 .
The methods described in FIGS. 8-10 (e.g., inferences made from a single sensor or a combination of sensors) may be used arrive at specific inferences of an animal's health or wellness. For example, the analysis of one or more sensors Nm may allow episodic and/or longitudinal inferences to be made regarding animal's health and wellness. As an example episodic inference that may be made using one or more sensors, in one embodiment a GPS geo-zone alert may be confirmed or canceled using, e.g., GPS sensor (as one example of the sensor provided on wearable device 101 ). Specifically, a geo-zone alert may be prone to false positives due to, e.g., temporary loss of communication with one or more satellites (which may thus be interpreted as movement of animal 401 ). However, in some embodiments, a GPS geo-zone alert may be compared with an accelerometer reading to corroborate/confirm the alert. Specifically, if the animal 401 is not moving (as determined from data received from the accelerometer) the geo-zone alert may be canceled.
Similarly, in some embodiments signal strength of, e.g., an RF signal may be compared to GPS position of animal 401 to confirm, e.g., a breach of a geo-zone. Specifically, a reading from the GPS may be indicative that the animal 401 has moved outside a geo-zone. However, if signal strength of an RF signal from a base station (received at RF antenna) is still rather strong, the GPS readings may be interpreted as a false positive (e.g., the result of losing communication with one or more satellites) and thus the alert may be canceled.
As another example episodic inference that may be made using one or more sensors, a reading of high acceleration (from, e.g., an accelerometer) may trigger additional sensors and/or otherwise be compared with data from additional sensors to determine if animal 401 was involved in an impact event (e.g., being hit by a vehicle). For example, a reading of high acceleration from the accelerometer may be supplemented with a reading from, e.g., a light meter and or a microphone on wearable device 101 (as two examples of internal sensors). If, in addition to the high acceleration reading, the wearable device received a high light incidence reading (e.g., headlights) and/or a high noise reading (e.g., impact) then an alert of a possible impact event may be returned.
As another example episodic inference that may be made using one or more sensors, a breach of a perimeter fence (as determined by RF antenna 109 , Wi-Fi, Bluetooth, or other RF technology) may be compared to readings from an ambient light, sound, temperature, and/or humidity sensor on wearable device 101 (as examples of internal sensors) to determine if animal 401 has in fact, e.g., left a house. If the sensed humidity, temperature, light, etc., is indicative of the animal 401 being outside, then the perimeter fence alert may be returned. However, if each reading is indicative of the animal 401 being inside, the breach of perimeter fence alert may be interpreted as a false positive and thus canceled.
As another example episodic inference that may be made using one or more sensors, data from, e.g., a microphone (as one example of a sensor) may be compared with reading from an accelerometer (as another example of a sensor) to determine if animal 401 has been, e.g., barking longer than a threshold period of time. For example, a reading from a microphone may be indicative of animal 401 barking, or may be due to some other event (e.g., thunder). However, data received from the accelerometer may confirm/negate that the animal has been barking according to whether or not a signature head movement or vibration of a barking event was sensed or not.
Further, sensed data from an inward looking antenna (e.g., a UWB antenna) may be compared with a microphone to form many inferences related to respiration quality and the like. For example, UWB antenna may be used to form an inference of animal's respiration quality by monitoring movement of muscles in the neck area (e.g., the muscles surrounding the animal's trachea 511 ). Further, the sensed UWB data may be corroborated with a microphone located on wearable device 101 and/or an external microphone (e.g., a microphone located on an owner's personal mobile device such as a smartphone, etc.) to make an inference regarding whether the animal 401 has kennel cough, bronchitis, etc.
As another example episodic inference that may be made using one or more sensors, noninvasive cardio output may be determined by measuring both heart rate (beats per minute), quality (fluctuations over the minute), and stroke volume to provide cardiac output using UWB technology on either an episodic or trending basis. Other derived conclusions from these measurements may also include a change in blood pressure over time and whether the animal is losing blood volume due external or internal bleeding. These sensors may be placed on the animal's chest near the sternum, at the front of the neck near the wind pipe and carotid arteries, or on other parts of the animal to pick up specific signals of interest.
As another example episodic inference that may be made using one or more sensors, noninvasive core temperature may be measured and/or derived from several internal and ambient thermistors. Further, microwave radiometry/thermometry (using a microwave antenna) along with other techniques may be used to determine fluctuations in core temperature which may be indications of hypothermia, hyperthermia, bacterial or viral infections, inflammation, on set of disease, immune-mediated or neoplastic diseases, extreme exercise, or ovulation.
As another example of an episodic inference that may be made using one or more sensors, noninvasive measurement of blockages in the digestive track can be accomplished by moving the wearable device 101 to the area of concern to allow readings and an upload of data from this activity using the UWB technology.
As another example episodic inference that may be made using one or more sensors, noninvasive measurement of the animal's drinking and eating habits may be measured independently or corroborated with other sensors using UWB technology by examining signals from the neck area including the esophagus and surrounding tissues.
In some embodiments, a base line measurement of animal 401 may be determined and then compared to subsequent data collection to determine, e.g., one or more of the inferences discussed herein. In some embodiments, data received from two or more sensors may be used to determine, e.g., that it is an appropriate time to collect this baseline data. For example, in some embodiments, a clock or other component (e.g., light meter, etc.) may be accessed to determine, e.g., that it is night time. Further, data from the accelerometer may be referenced to confirm that, e.g., animal 401 is sleeping (as indicated by no or little acceleration). In such embodiments, a baseline measurement of one or more vital signs and/or physiological signs may be taken in response to the one or more sensors indicating that animal 401 is sleeping.
The above methods of determining episodic inferences from one or more sensors may be more readily understood with reference to a specific example. In one embodiment, wearable device 101 may include an accelerometer, a microphone (as examples of internal sensors) and/or cardiopulmonary sensors (e.g., UWB device). In such an embodiment, the accelerometer may measure a high acceleration event, and the wearable device 101 /DMS 301 may interpret the acceleration as indicative of a possible impact event (e.g., the animal 401 was hit by a vehicle). The wearable device 101 /DMS 301 may then corroborate or confirm this interpretation by referencing other sensors, e.g., microphone. For example, if the microphone sensed a loud noise at the moment of the high acceleration, the inference of an impact event may be confirmed. This may then trigger other sensors, such as cardiopulmonary sensors (e.g., UWB device). For example, the cardiopulmonary sensors may check animal 401 for anomalies, which may include, e.g., checking animal 401 for loss of blood volume (indicative of, e.g., internal or external bleeding).
The example of an episodic inference of an impact event made by the wearable device 101 and/or DMS 301 is illustrated in FIG. 11 . FIG. 11 illustrates how readings of one or more sensors may be interpreted as indicating that an event has occurred. As it shown in FIG. 11 , signals from five sensors are used with the sensors identified as Na, Nb, Nc, Nd, and Ne, respectively. The readings from sensors Na 1101 , Nb 1102 , and Nc 1103 are weighted independently by weighting factors W Na 1104 , W Nb 1105 , and W Nc 1106 , respectively. Next, in step 1107 , it is determined if the weighted combination of the readings of these three sensors is above a threshold al. If no, then the system ignores the sensor readings in step 1108 and returns to monitoring the animal. If yes, then this determination is stored in step 1109 and the alert provided as alert level 1 in step 1110 .
FIG. 11 also includes the ability for determination of a second alert level (alert level 2 ). For instance, the system knows after step 1107 that alert level 1 has been reached. The system may additionally check in step 1111 the weighted combination or perform an additional weighting and compare the weighted combination against a second alert level threshold, here, the a 2 threshold. If yes from step 1111 , that is second alert level a 2 is stored in step 1112 and alert level 2 is identified to the owner/DMS in step 1113 .
If no from step 1111 as having not found a second alert level based on the initial weighted sensor readings from sensors Na, Nb, and Nc, there may be additional sensor inputs that allow a determination that the second alert level has been reached. For instance, sensor readings from sensors Nd 1114 and Ne 1115 may be obtained. For the sensor reading from sensor Nd, the system determines in step 1115 if the sensor reading is below a low threshold for sensor Nd. If yes, then this determination is stored in step 1112 and the alert level 2 is provided in step 1113 . If no from step 1115 , the system determines in step 1116 if the sensor reading is above a high threshold for sensor Nd. If yes, then this determination is stored in step 1112 and the alert level 2 is provided in step 1113 . If no from step 1116 , then the system continues to provide the alert level 1 in step 1110 .
A similar determination may be made for reading from sensor Ne. For the sensor reading from sensor Ne, the system determines in step 1118 if the sensor reading is below a low threshold for sensor Ne. If yes, then this determination is stored in step 1112 and the alert level 2 is provided in step 1113 . If no from step 1118 , the system determines in step 1119 if the sensor reading is above a high threshold for sensor Ne. If yes, then this determination is stored in step 1112 and the alert level 2 is provided in step 1113 . If no from step 1119 , then the system continues to provide the alert level 1 in step 1110 .
Finally, one of the original sensor levels may be reviewed to determine if it is outside of a profile for that sensor. For instance, in step 1120 , the sensor readings of sensor Nc are compared against a profile for that sensor. If the readings are outside of that profile, then this determination is stored in step 1112 and the alert level 2 is provided in step 1113 . If no from step 1120 , then the system continues to provide the alert level 1 in step 1110 .
The following explains how FIG. 11 may be applied to specific sensor readings to determine if an event has occurred. The following example explains how a determination is made that a high impact event has occurred. Here, sensors Na, Nb, Nc, Nd, and Ne are represented by a light meter sensor n 1 , a microphone/peak sound sensor n 2 , an accelerometer n 3 , a GPS receiver n 4 , and a cardiopulmonary sensor n 5 , respectively.
At step 1103 , accelerometer (n 3 ) senses a high acceleration event (e.g., 10+G's) potentially indicative of a high-impact event. In this embodiment, the accelerometer (n 3 ) acts as a “master” sensor such that when it has sensed this episodic condition at step 1103 (e.g., high accelerations possibly indicative of an impact event), it may control the sensing and/or data reporting of other sensors to confirm/corroborate the event. Specifically, processor 101 may use the high signal on accelerometer n 3 to look back for recent readings from light meter n 1 and microphone n 2 . Those recent readings may have been stored in storage 105 or in storage 119 , depending on the sensor. The effect is that accelerometer sensor n 3 is, for this instance, a master sensor and the light meter n 1 and microphone n 2 are the slave sensors.
The previous readings from the slave sensors are reviewed to look for episodic threshold events to create a more accurate picture as to what has transpired over the previous time interval and possibly confirm a possible high impact event from accelerometer n 3 . Thus, at step 1105 processor 100 retrieves stored data from the microphone/peak sound sensor (n 2 ) for a time period immediately preceding and overlapping with the high acceleration reading, and at step 1107 processor 100 retrieves stored data from the light meter n 1 for a time period immediately preceding and overlapping with the high acceleration reading.
At steps 1104 - 1106 , the data received from each sensor may be weighted and combined into a single result to determine in step 1107 if the constructed profile meets a high degree of probability that an event of interest (e.g., impact) has occurred. For example, if the light meter (n 1 ) sensed a high incidence of light (potentially indicative of headlights), and/or if the microphone/peak sound sensor (n 2 ) sensed a loud noise (potentially indicative of a being impacted by a vehicle), then the method may determine at step 1107 that an impact has in fact occurred. If the other readings do not confirm the possible impact event, then the data may be ignored at step 1108 . Regardless, the data received may be written and/or stored locally at step 1109 for subsequent upload to the DMS 301 .
If the combined and corroborated data meets certain conditions (e.g., each is indicative of an impact event) in step 1107 , the master sensor (in the depicted embodiment, accelerometer n 3 ) may trigger and/or change states other sensors (including itself) in order to, e.g., take individual spot readings, schedule-based readings, or change each sensor's sensing configurations. If the readings are inconclusive, the sensors are instructed to continue reading.
For example, in the depicted embodiment, at step 1109 , the accelerometer (n 3 ) changes (as being controlled by processor 100 ) from being in an interrupt mode (e.g., looking for episodic events) to a real-time monitoring of motion activities. This real-time monitoring may be compared to a profile to determine if the animal's gait has changed dramatically as determined in step 1120 . At step 1117 , the GPS sensor (n 4 ) is instructed (i.e., controlled by processor 100 ) to determine location, speed, and/or direction of the animal 401 . If the animal 401 is moving in a sustained fashion, this reading would have a lower risk ratio assigned to it. Further, at step 1107 , the cardiopulmonary sensor (n 5 ) may be triggered to check on heart rate, respiration rate, stroke volume, and/or a change in blood pressure. The cardiopulmonary sensor (n 5 ) may thus look for anomalies (e.g., loss of blood) and assign a risk ratio to the readings. Or, in other words, the processor 100 may look for anomalous readings from the cardiopulmonary sensor n 5 and assign a risk ratio to those readings.
At steps 1115 , 1116 , 1118 , and 1119 , the processor in the wearable device 101 and/or DMS 301 may compare the data from one or more of the above sensors to determine, e.g., an alert level following the determined episode (e.g., impact event). For example, after considering all of the above weighted data points, the processor may determine that the event recorded merits various levels of alerts (at steps 1110 and 1113 ) to be sent to the owner and/or the veterinarian based on the reliability of the sensor readings. Further, the wearable device 101 may be instructed to continue reading at steps 1110 and 1113 in order to continually monitor the animal's progress following the impact event.
The following equations describe the weighting of the values of the sensors and the comparison against the alert level thresholds. Equation (1) below describes how a sensor reading from sensor Nc is checked against the threshold for sensor Nc:
If ( n c >n c threshold ),then alert for n c exceeding n c threshold (1)

Equation (2) below describes how a sensor reading from sensor Nc is checked against the threshold for sensor Nc and, if the threshold is exceeded, then determining if a weighted combination of sensor readings Na and Nb and Nc exceed the alert level 1 threshold:













If






(


n
c

>

n

c





threshold



)


,
then
,








if







(


n
a


max





over





time





T





1

)


n

a





threshold




x






w
a


+



(


n
b


max





over





time





T





2

)


n

b





threshold




x






w
b


+



(


n
c


max





over





time





T





3

)


n

c





threshold




x






w
c



≥

a
1


,










then





alert





for





alert





1






(
2
)







where:
a 1 is the alert level 1 threshold such that a value above a 1 results in alert level 1 while a value below a 1 does not result in an alert;
Times T 1 , T 2 , and T 3 are the time intervals in which the previous readings for sensors Na, Nb, and Nc are reviewed; and
Wa, Wb, and Wc are the weighting values for each of the Na, Nb, and Nc sensor readings.
Notably, equation (2) normalizes the values of each sensor by dividing the max value of the sensor during a time window (or min as appropriate) by the threshold. This permits the individual units of each sensor to cancel out. Next, the weighting factors scale each normalized sensor reading such that they can be added and compared against the threshold for alert level 1 (a 1 ).
Equation (3) below describes a similar analysis as that of equation (2) but sets the alert level threshold at the alert level 2 a 2 threshold:













If






(


n
c

>

n

c





threshold



)


,
then
,








if







(


n
a


max





over





time





T





1

)


n

a





threshold




x






w
a


+



(


n
b


max





over





time





T





2

)


n

b





threshold




x






w
b


+



(


n
c


max





over





time





T





3

)


n

c





threshold




x






w
c



≥

a
2


,










then





alert





for





alert





2






(
3
)







where:
a 2 is the alert level 2 threshold such that a value above a 2 results in alert level 2 while a value below a 2 does not result in an alert;
Times T 1 , T 2 , and T 3 are the time intervals in which the previous readings for sensors Na, Nb, and Nc are reviewed; and
Wa, Wb, and Wc are the weighting values for each of the Na, Nb, and Nc sensor readings.
Equation (4a) and (4b) relate to equation (2) but also includes the slave sensor analyses of FIG. 11 :













If






(
master
)



(


n
a

>

n

a





threshold



)






and












(


n
a


max





over





time





T





1

)


n

a





threshold




x






w
a


+



(


n
b


max





over





time





T





2

)


n

b





threshold




x






w
b


+



(


n
c


max





over





time





T





3

)


n

c





threshold




x






w
c



≥

a
1







(

4

a

)







then activate slave (4b)
If

((( n d <n d low threshold ) or ( n d >n d high threshold ))
or
(( n e <n e low threshold ) or ( n e >n e high threshold ))
or
(( n a #preexisting profile for n a )), then alert level 2 , otherwise alert level 1 .

where:

a 1 is the alert level 1 threshold such that a value above a 1 results in alert level 1 while a value below a 1 does not result in an alert;
Times T 1 , T 2 , and T 3 are the time intervals in which the previous readings for sensors Na, Nb, and Nc are reviewed;
Wa, Wb, and Wc are the weighting values for each of the Na, Nb, and Nc sensor readings; and
“preexisting profile for n a ” is a profile for expected values of n a over a time interval.
Here, alert level 2 is defined by being activated by both master and slave reaching predefined levels. Alert level 1 is defined by being activated by only the master reaching its predefined level but the slave not reaching its predefined level.
The equations above also permit the sensors to be located on other devices based on the time T being evaluated for each sensor reading. So, once a common time is determined (for instance, the time T(Nc) at which the reading from sensor Nc exceeded the Nc threshold), the other sensor readings are time normalized from that time T(Nc) and evaluated.
Sensors Located on Different Devices
As described above, all of the sensors may be located on wearable device 101 or some located on the wearable device 101 and others located on a separate device. A separate device may be a user's smartphone (e.g. the microphone on the smartphone). In short, data may be captured and compared from sensors located on more than one device (e.g., wearable device 101 and a user's mobile device) and compared to determine, e.g., an episodic inference about the animal's health and wellness. For example, FIG. 12 illustrates one example method for capturing sensor data from more than one device which can then be forwarded to the DMS 301 and analyzed to determine an inference regarding animal's health and wellness (in the depicted example, respiration inferences). As with FIG. 11 , the timeline 12011 of FIG. 12 indicates a relative time that each step is performed relative to one another. In FIG. 12 , at step 1201 a user opens a mobile device application. For example, the health-monitoring system as described herein may include a companion mobile application that can be downloaded to an animal 401 owner's smartphone, tablet, computer, etc., which may capable of triggering sensors on demand. A user may be the animal's owner or a veterinarian, etc. In step 1202 , the user may select a function they wish to collect data about. The specific sensors selected for capturing and returning data may vary depending on what particular inference, etc., the user triggers. In the embodiment depicted in FIG. 12 , the user selects respiration analysis. At step 1203 , commands may be sent to the sensors to collect and/or forward data related to this respiration analysis. For example, because the user selected “respiration analysis,” a command may be sent to a cardiopulmonary sensor (n 5 ) and to an accelerometer (n 3 ), both located on wearable device 101 , and to a microphone (n 14 ) located on the user's mobile device. At steps 1204 , 1205 , and 1206 , each respective device may collect data and/or retrieve previously collected data. These sensors could be placed on standby and triggered based on the start of an event (as, for instance, a coughing fit).
In the following three examples, the following scenarios are explained: no triggering between the mobile device and the wearable device (only being synced by the DMS), triggering of the mobile device to start recording by the wearable device, and triggering of the wearable device to start recording by the mobile device. In the first example, an application executing on the user's mobile device may be executing and recording audio files with time stamping. The DMS may correlate the audio file with readings from accelerometers based on time-stamps of data obtained from the accelerometers. In the second example, the mobile device or the wearable device may trigger the other based on sensed levels exceeding a threshold. For instance, the mobile device may be waiting for the wearable device to indicate that the wearable device's accelerometer has started sensing the coughing fit at which point the wearable device alerts the mobile device. In response to the alert, the mobile device may start recording an audio file with time stamps. In this example, the excess, uninteresting audio file recorded before the dog started coughing is not recorded. In the third example, the mobile device informs the wearable device that the microphone on the mobile device has picked up the sounds of the coughing fit and that the wearable device is to monitor the animal. In the following three examples, the following scenarios are explained:
Each piece of collected data at steps 1204 - 1206 may be time-stamped such that, when analyzed, each may be lined up in order or otherwise synchronized to correctly aggregate and consider each piece of data with the others. At step 1207 , the data collected on wearable device 101 is uploaded to the DMS 301 , and at step 1208 , the data collected at the user's mobile device is uploaded to DMS 301 . At step 1209 , the uploaded data are correlated against each other based on synchronizing the timestamps to determine when a relevant. Of coughing has begun. Next, in step 1210 the data are analyzed at the DMS 301 to determine appropriate inferences regarding the animal's health and wellness (in the depicted example, respiration quality).
For example, the combined data may lead to an inference that the animal 401 is suffering from kennel cough or bronchitis. Further, because in some embodiments the data will be time-stamped, an inference may be readily determined even though the sensor readings are coming from disparate sources (here, wearable device 101 and a mobile device). Although as described the analysis step 1210 is performed at the DMS 301 , in other embodiments the analysis may be performed at the user's mobile device and/or the wearable device 101 .
In addition to episodic inferences made using the methods depicted in FIGS. 8-12 , longitudinal inferences (e.g., trending inferences) may be made using the above described methods. That is, because collected data may be stored locally in the wearable device (at, e.g., steps 805 , 907 , 1005 / 1013 , and/or 1109 / 1112 ) and/or uploaded to the DMS 301 for storage, changes or fluctuations, etc., in data over time may be monitored, and according longitudinal (trending) inferences may be made regarding animal's health and wellness.
By way of example, in some embodiments animal's long-term weight fluctuations may be monitored and inferences may be made about the animal 401 accordingly. For example, monitoring long-term weight fluctuations are important as a lean pet has a 15% increase in lifespan (+2 years) and may also be a precursor to other developing conditions. On the other end of the scale, rapid weight loss may be indicative of a digestive track blockage or cachexia where the body is breaking down protein and fat due to the onset of diabetes. Thus, by monitoring and comparing an animal's weight overtime, an inference as to the animal's health and wellness may be determined.
As another example of a longitudinal inference that may be determined using one or more sensors, an activity level of an animal may be monitored (using, e.g., an accelerometer, GPS, etc.). Further, the measured activity levels may be adjusted by the DMS 301 for weekends and weekday lifestyle profiles of the animal 401 and/or the animal's owner. For instance, if the owner takes the animal for walks at 3 am, this may be identified by the owner to the DMS and the DMS refrain from alerting the owner that the animal has left the owner's house at night. Inferences made from the monitored activity levels may indicate that the animal is not being provided with enough exercise opportunity or that conditions such arthritis are slowing the animal down during times of self-initiated activity.
As another example of a longitudinal inference that may be determined using one or more sensors, the animal's eating and hydration habits may be monitored over time. Hydration and eating fluctuations may be important indicators of developing polyphagia and polydipsia conditions related to diabetes.
As another example of a longitudinal inference that may be determined using one or more sensors, sleep patterns of an animal may be monitored to form inferences regarding animal's health and wellness. Sleep patterns may be important indicators of underlying issues with pets such as osteoarthritis. Some owners may assume that an animal sleeping more is just a result of old age, whereas, in reality, it may be an indicator of developing medical conditions. For example, an animal may not limp or whine when excited during play and act like a younger dog but will pay for it later. This may manifest itself in longer rests, stiffness on rising, and resistance to go on their regular walks. Other reasons for longer sleep periods could be caused by thyroid, kidney, or liver disease. Animals may also have sleep disruption caused by obsessive-compulsive behavior disorders. In some embodiments, sleep patterns may be derived by the DMS 301 and collaborated with owner personal observations 312 .
According to other aspects of the disclosure, longitudinal inferences may be determined using the provided UWB technology of the wearable device (e.g., using UWB device). For example, in one embodiment respiration monitoring may uncover abnormal signs such as panting while resting, using more abdominal muscles to breath, labored breathing, asymmetrical breathing, increased or decreased breathing rates, wheezing, coughing, and choking.
As another example of a longitudinal inference that may be determined using UWB technology, animal's heart rate may be monitored over time by UWB device. Heart rate monitoring may uncover increased or decreased heart rate and/or abnormal rhythms, which may include the heart speeding up and slowing down or missing beats. In additional embodiments, stroke volume measured overtime may be used to derive the overall fitness level of the animal 401 and/or indicate that the animal 401 is developing conditions that would cause it to be lower.
As another example of a longitudinal inference that may be determined using UWB technology, an animal's blood pressure changes (both increased and decreased blood pressure) may be monitored. Blood pressure changes from a base line (which may be measured, e.g., when animal 401 is sleeping or otherwise in a state of low activity as discussed) may be an indicator of hypertension developing which may lead to other severe medical conditions.
In any of the above embodiments, collected data may be time-stamped in order determine time-dependent inferences. That is, time stamping the various sensing activities and the ability to look backward in time allows for a root-cause analysis to determine an adverse event (e.g. the animal was walking fine, but then played fetch and is now limping). Further, in some embodiments, time-stamping may also allow for the analysis of the rate of change which in turn can be used to predict a possible outcome (e.g. the animal is running at an increasing rate of speed towards the outer area of the geo-zone and thus is likely to breach that zone).
FIG. 13 presents a table 1301 summarizing illustrative attributes of some sensors that may be located on wearable device 101 or located external to wearable device 101 and used in conjunction with the health-monitoring system described herein according to some aspects of the disclosure. Specifically, 1301 contains column 1303 denoting a number of each sensor (denoted as Nm), column 1305 indicating the type of each sensor, column 1306 describing the location of the sensor relative to the wearable device, column 1307 indicating a primary purpose of each sensor, column 1308 describing a general category of sensor, column 1309 indicating whether each sensor may act as a master or a slave sensor (as described herein with respect to FIG. 14 ), column 1311 indicating a secondary purpose (if any) of each sensor.
By way of example, in this embodiment N 1 refers to a light meter and/or spectrometer located on wearable device 101 . As denoted in column 1307 , the light meter's primary purpose may be to monitor light levels surrounding wearable device 101 (and thus animal 401 ). Further, as indicated in column 1309 , the light meter may only act as a slave sensor and thus, in this embodiment, may not control other sensors. As indicated in column 1311 , the light meter may also have a secondary purpose, here serving as an indoor/outdoor indicator (by, e.g., sensing UV levels) or analyzing nearby chemical signatures in the air.
FIG. 14 presents a table indicating illustrative master/slave relationships of each sensor presented in FIG. 13 according to or more embodiments of the disclosure. Specifically, FIG. 14 includes rows identifying each sensor as well as columns identifying each sensor. The values in each cell identify the relationship as a row sensor is a master sensor in contrast to the slave identified in the column sensor where the intersecting cell includes an “X”. At the intersection of the same sensor in the row and column title, the cell value is identified by “I” to indicate if the identical sensor. Interestingly, in some implementations, each sensor may act as a master to itself (e.g., control further collection of data by itself in response to a sensed reading). An example of this is shown in step 1120 of FIG. 11 identifying whether the readings from sensor Nc are outside of an expected profile.
By way of example, as indicated by each “X” or darkened cell in the row following “N 3 ” listed, in some embodiments accelerometer (N 3 ) may act as a master to slave sensors N 1 (light meter), N 2 (peak sound), N 3 (itself, accelerometer), N 4 (GPS), N 5 (cardiopulmonary), N 6 (temperature), N 8 (Wi-Fi), N 9 (Bluetooth), N 10 (RF), and N 11 (GSM). Further, as indicated by each “X” or darkened cell in the column below “N 3 ”, in some embodiments accelerometer (N 3 ) may serve as a slave to other master sensors, namely N 3 (itself, accelerometer), N 5 (cardiopulmonary), N 13 (battery strength), and N 14 (mobile microphone).
FIG. 15 relates to various operation modes and how each sensor may operate in the various operation modes. Column 1501 identifies the sensor by number. Column 1502 identifies a sensor type. Column 1503 identifies how each sensor operates in a profile operation mode. Column 1504 identifies how each sensor operates in an airplane (no RF radios operative) operation mode. Column 1505 identifies how each sensor operates in a location alert operation mode.
For instance, FIG. 15 identifies the peak sound sensor, the accelerometer, and the time of day sensor (e.g., an internal clock) are not affected by the specific profile settings when in the profile mode as shown in column 1503 . The remaining sensors may have different operations based on the profile.
In the airplane operation mode 1504 , most of the sensors are off while peak sound is in a standby state the accelerometer, the ambient temperature sensor, and the time of day sensor are on. In other words, the operation of the sensors in the airplane mode identifies that all radios, sensors, and/or components that generate significant that generate significant electro-magnetic radiation are disabled.
In the location alert operation mode 1505 , all sensors that may help determine the location of an animal are on, including light meter, accelerometer, GPS, WiFi signal detector, Bluetooth signal detector, RF signal detector, and GSM signal detector sensors. The remaining sensors may be turned off to help conserve power. The battery strength sensor may also be left on in the location alert mode 1505 to identify to the collar when it is running low on power. For example, the cardiopulmonary sensor n 5 is disabled in favor of the GPS sensor/radio n 4 , the Wi-Fi sensor/radio n 8 , the Bluetooth sensor/radio n 9 , the RF sensor/radio, n 10 , and the GSM sensor/radio n 11 , depending on which of these sensors/radios are present.
FIGS. 16A-16G relate to different profiles usable by wearable device 101 . In each of FIGS. 16A-16G , column 1601 identifies the sensor number and columns 1602 identifies the sensor type.
FIG. 16A describes a first profile, Profile 0 , which relates to a normal monitoring profile set by the owner. The profile type identified in cell 1603 A and its title identified in cell 1604 A. Here, the range between the low threshold 1605 A and the high threshold 1606 A is set relatively large, the frequency of operation of each sensor is relatively infrequently, and granularity for the readings of various sensors is low. This profile is an example of a normal profile set by the owner. For instance, a processor operating under Profile 0 of FIG. 16A has a low granularity for accelerometer sensor n 3 . The low granularity may take the form of a low pass filter applied to a signal from the accelerometer sensor n 3 . The low pass filter may smooth any instantaneous accelerometer output level to eliminate and/or reduce the triggering of the accelerometer high threshold when the instantaneous output is above the high threshold but while the average output is low. Alternatively, the low pass filter may be replaced with a smoothing filter (e.g., a convolution filter with a longer time constant) to reduce any errant spikes in the signal from the accelerometer n 3 . Further, the above described filters may be part of the processor such that the processor ignores or is less sensitive to acceleration spikes with short duration
FIG. 16B describes a second profile, Profile 1 , which relates to an enhanced monitoring profile set by the owner. The profile type identified in cell 1603 B and its title identified in cell 1604 B. Here, the range between the low threshold 1605 B and the high threshold 1606 B is narrow compared to that of Profile 0 of FIG. 16A , the frequency of operation of each sensor is relatively more frequent, and granularity for the readings of various sensors is high. This profile is an example of an enhanced profile where the owner is concerned about the pet's current health and desires more information to be obtained by the collar. In contrast to the Profile 0 of FIG. 16A , this Profile 1 has enhanced sensitivity as shown in some of the trigger point for the low thresholds of column 1605 B being higher and the trigger point for the high thresholds of column 1606 B being lower. Also in some instances, the frequency of monitoring in column 1601 B is more often. Similarly, the granularity as shown in column 1608 B is also high. For instance, for accelerometer n 3 , the granularity is described in column 1608 B as being high. With respect to the example of the low pass filter, the filter may be removed or modified to reduce the level of filtering of higher frequency signals. With respect to the example of the smoothing filter, the time constant (or window of time over which the smoothing takes place) is reduced to permit higher frequency acceleration signals to be analyzed by a processor. Also, as described with respect to FIG. 16A , the filters may be part of the processor such that the processor adjusts internally how sensitive it is to the outputs of various sensors based on a current profile.
FIG. 16C describes a third profile, Profile 2 , which relates to a normal monitoring profile set by the veterinarian. The profile type identified in cell 1603 C and its title identified in cell 1604 C. Here, the range between the low threshold 1605 C and the high threshold 1606 C is set relatively large with even some sensors not being used as the veterinarian may not need the readings from the sensors, the frequency of operation of each sensor is relatively infrequently, and granularity for the readings of various sensors is low. This is an example of a profile where the vet may be monitoring the pet's current health to establish a baseline or as a function of general monitoring (for example, in preparation for a checkup).
FIG. 16D describes a fourth profile, Profile 3 , which relates to an enhanced monitoring profile set by the veterinarian. The profile type identified in cell 1603 D and its title identified in cell 1604 D. Here, the range between the low threshold 1605 D and the high threshold 1606 D is set relatively narrow, the frequency of operation of each sensor is relatively frequent, and granularity for the readings of various sensors is high. Again here, some sensors are disabled as the veterinarian may have no need for the readings from those sensors. For instance, this profile may be used before surgery or a procedure (e.g., teeth cleaning with the animal being anesthetized) is performed on the animal to ensure no recent dramatic events have occurred to the animal prior to the surgery/procedure.
For instance, this profile may be used after surgery or after a procedure to monitor for possibility of complications arising from the surgery. Based on the level of need for monitoring the animal, the rate at which information is provided to the veterinarian may be further modified in accordance with the examples of FIG. 22 as relating to the following:
A. Identification of events by the wearable device and uploading those events to the veterinarian, B. Logging of raw data from the sensors and batch uploads of the logged data to the veterinarian, or C. Continuous uploads of raw data to the veterinarian.

With respect to the above description and the description of FIG. 22 , the uploads of the identified events and/or raw data to the veterinarian may be a direct transfer from the wearable device to a remote device (for instance, to a computer on a same local Wi-Fi network as the wearable device) or may be an indirect transfer from the wearable device to the DMS which then forwards to the veterinarian (or makes available for the veterinarian to access) the identified events and/or raw data from the wearable device. Further, the DMS may further derived events from the raw data and possibly the device-derived events from the wearable device. These DMS-derived events may be further provided to the veterinarian or made available for viewing by the veterinarian as desired.
FIG. 16E describes a fifth profile, Profile 4 , which relates to a monitoring profile for a first specific symptom type as set by the veterinarian. The profile type identified in cell 1603 E and its title identified in cell 1604 E. Here, the range between the low threshold 1605 E and the high threshold 1606 E is set relatively narrow, the frequency of operation of each sensor is relatively frequent, and granularity for the readings of various sensors is high for some sensors but low for others. In this profile, the veterinarian is concentrating on values from some sensors over other sensors. For instance, the veterinarian may be monitoring for gait-related issues based on the accelerometer frequency sampling being “always on” and the granularity being “high”.
FIG. 16F describes a sixth profile, Profile 5 , which relates to a monitoring profile for a second specific symptom type as set by the veterinarian. The profile type identified in cell 1603 F and its title identified in cell 1604 F. Here, the range between the low threshold 1605 F and the high threshold 1606 F is set relatively narrow, the frequency of operation of each sensor is relatively frequent, and granularity for the readings of various sensors is high for some sensors but low for others. In this profile in contrast to that of Profile 4 , the veterinarian is concentrating on values from a difference of sensors then important sensors of Profile 4 of FIG. 16E . Here, the veterinarian may be monitoring for a cardiopulmonary-type symptoms or similar set of symptoms by the cardiopulmonary sensor n 5 frequency being set to obtain a reading every minute with its granularity set to high.
FIG. 16G describes a seventh profile, Profile 6 , which relates to an enhanced monitoring profile set by the veterinarian in which some sensors are operated continuously as opposed to their standard intermittent usage. The profile type identified in cell 1603 G and its title identified in cell 1604 G. Here, the range between the low threshold 1605 A and the high threshold 1606 A is set relatively arrow, the frequency of operation of each sensor depends on its importance. For those sensors that are not important, they are not operated and in contrast other sensors are operated continuously. For instance, this profile may be used when an animal is recovering from surgery and the veterinarian desires continuous readings of the vital signs/physiological signs of the animal without stressing the animal by having individual sensors for each vital sign/physiological sign being separately attached. Alternatively, this profile may be used when the animal is in critical condition and is in a constantly monitored state. In this profile, some items are not monitored as they are not relevant when staying in hospital. For instance, monitoring the ambient temperature via sensor n 6 or monitoring for GPS signals with sensor n 4 are not needed. This profile of FIG. 16G enables veterinarians to use the wearable device 101 in place of separately attached individual sensors that would normally be attached individually to the animal.
FIG. 18 shows an example of how various sensor profiles may be modified based on breed information of the animal to which the monitoring devices attached in accordance with one or more aspects of the disclosure. Specifically, column 1801 identifies those sensors which may be modified or adjusted in sensitivity when processing based on the type of breed of animal. For instance, high and low thresholds for cardiopulmonary sensor n 5 may be adjusted upwards for a breed that has a high average heart rate and downwards for a breed that has a low average heart rate.
FIG. 18 shows an embodiment with different operation modes of the wearable device in accordance with one or more aspects of the disclosure. In this embodiment, the wearable device operates in one of three operation modes: a profile mode 1802 , an airplane mode 1803 , and a location alert mode 1804 . The collection of operation modes is shown as group 1801 and the collection of profiles are shown as group 1802 . In this embodiment, two profiles may be implemented in the wearable device: owner profile 1805 and veterinarian/third-party profile 1806 . Based on the selection of the operation mode, wearable device 1807 operates as designated by the particulars of the operation mode. Finally, based on the designation in the operation mode of what and when to upload content to the remote data management system, the wearable device 1807 uploads content in accordance with the operation mode.
For instance, in the profile operation mode 1802 , this operation mode (and optionally the specific profile) identifies that content from the wearable device 1807 is to be uploaded to the remote data management system 1808 in batches. Next, in the airplane operation mode 1803 , as all radio transmission functions are disabled while in the airplane operation mode 1803 , the content collected while in operation mode 1803 is stored in wearable device 1807 and subsequently uploaded to remote data management system 1808 only when switched out of airplane mode 1803 . Further, when operating in the location alert operation mode 1804 , content information is uploaded to the remote data management system 1808 . For instance, in one example where the owner is attempting to locate the animal as soon as possible, the location content may be uploaded on a continuous basis to the remote data management system 1808 . The data uploaded from the wearable device may include location information from a GPS receiver sensor and/or triangulation information from received cell tower signal strengths and/or IP addresses of Wi-Fi access points, merely storing a list of time stamped IP addresses, or the like. The uploading of data may be real-time or may be batched. With respect to monitoring Wi-Fi access points, the wearable device 101 may keep track of the various access points encountered over time and upload a list of those access points so as to provide a list of locations (or approximate locations) visited throughout the day (or other interval) (thereby providing breadcrumb information of where the wearable device has been throughout the day).
FIGS. 19A-19B show the order in which operation modes take precedence over profiles based on the embodiment of FIG. 18 in accordance with one or more aspects of the disclosure. As used in FIGS. 19A-19B , the “switches” can be hardware switches, software switches or a combination of both. A hardware switch may be a switch located locally on the wearable device that permits selection of one of the operation modes described in FIG. 18 . A software switch is a remotely operated command to the wearable device to shift into one of the operation modes of FIG. 18 and/or profiles. The software switch maybe operated by the owner, a veterinarian, and or a third party. For instance, airport personnel may be included in the group including the third-party where the airport personnel may be able to access the wearable device to set it into the airplane operation mode 1803 . The combination of hardware and software switches permits a device to respond to either a hardware switch operation (actual switch or a double tap of the device-sensed by the internal accelerometer) or a software switch operation. For instance, external hardware switches may be located at one or more locations on the wearable device 101 at, for instance, locations A-C on the wearable device 101 of FIG. 5 or as part of collar/harness 402 . Here, the hardware switches may be respective parts of clasp 505 at locations H and I and operated by locking together the parts of clasp 505 .
FIG. 19A shows a deprecated order in which an airplane mode switch 1901 has the highest level of precedence. Next, a location alert switch 1902 has the second-highest level precedence. Third, the lowest level of precedence is profiles in profile group 1903 including owner profile 1904 and veterinarian/third-party profile 1905 .
FIG. 19B shows the different operation modes based on operation of the switches of FIG. 19A . First, if the airplane mode switch is on, then the wearable device operates in the airplane mode 1907 . If the airplane mode switch is off 1906 , then the wearable device looks to the state of the location alert switch. If the location alert switch is on, then the wearable device operates in the location alert operation mode 1909 . If the location alert switch is off 1908 , then the wearable device operates in one of the profile modes 1910 (for instance, in the owner profile 1911 or the veterinarian/third-party profile 1912 ).
FIG. 20 shows an alternative embodiment with different profiles including profiles replacing the operation modes of the embodiment of FIG. 18 in accordance with one or more aspects of the disclosure. Profiles 2001 include airplane profile 2004 , location alert profile 2005 , owner profile 2002 , and veterinarian/third-party profile 2003 . The selected profile from profiles 2001 dictate how wearable device 2006 operates and uploads data to the remote data monitoring system 2007 (similar to the operation mode/profiles of FIG. 18 ).
FIGS. 21A-21B show the combination of different profiles of the embodiment of FIG. 20 with options of profile selection by one or more switches in accordance with one or more aspects of the disclosure. FIGS. 21A-21B described profiles being designated by hardware/software/combination switches (the switches having been described with respect to FIGS. 19A-19B ). In FIG. 21A , the collection of profiles 2101 includes owner profile 2102 , veterinarian/third-party profile 2103 , airplane mode profile 2104 , and location alert profile 2105 . FIG. 21B shows the collection of profiles 2110 with the airplane mode switch and the locations mode switch designating at least some of the profiles. For instance, when airplane mode switch 2112 is on, the wearable device operates in airplane mode profile 2113 . When airplane mode switch is off 2111 , the location alert switch status is checked. If the location alert switch is on 2115 , the wearable device operates in the location alert profile 2118 . If the location alert switch is off 2114 , the wearable device operates in one of the owner profile 2116 or the veterinarian/third-party profile 2117 (as separately designated by the owner and/or veterinarian/third-party).
FIG. 22 shows an example of how profiles may be selected in the wearable device as well as in the DMS in accordance with one or more aspects of the disclosure. Wearable device 2201 shown relative to DMS 2213 . At step 2202 , an initial profile is set for the wearable device 2201 . In step 2203 , it is determined whether a sensor or combination of sensors has exceeded one or more thresholds as described herein. If yes, then the wearable device modifies its own profile to change to a different profile or operation mode as shown in step 2204 . Also, as shown by the yes arrow extending down from step 2203 , the derived events may be uploaded to the DMS in step 2205 , raw data may be uploaded to the DMS in batches as shown in step 2206 , or raw data may be continuously uploaded to the DMS in step 2207 depending on the new profile or new operation mode. If no from step 2203 , the derived events may be uploaded to the DMS in step 2205 , raw data may be uploaded to the DMS in batches as shown in step 2206 , or raw data may be continuously uploaded to the DMS in step 2207 depending on the current profile or current operation mode.
Next, content from wearable device 2201 is received at the DMS 2213 at step 2208 . In step 2209 , the data is stored (for instance, in a database in one or more servers with dynamic or solid-state memory as shown by database 2210 ) and subsequently analyzed. If in step 2211 , an alert has been triggered from the analyzed data, then DMS 2213 instructs wearable device 2201 to change to a different profile or operation mode in accordance with the alert level determined in step 2211 . Alternatively, if no from step 2211 , no alert has been determined and the DMS 2213 continues to monitor for content from wearable device 2201 in step 2208 .
FIG. 23 shows an example of how output from various sensors may be stored for an interval of time and then discarded in accordance with one or more aspects of the disclosure. FIG. 23 shows the past history for signals from accelerometer 2301 , light sensor 2302 , and sound sensor (microphone) 2303 . In this example, older readings 2309 from accelerometer 2301 were below an accelerometer threshold level {Threshold(acc)}. However more recently, the signal from the accelerometer rose to level 2308 , which is above {Threshold(acc)}.
As described above, processor 100 may then evaluate previous readings from other sensors. Previous values from light sensor 2302 are evaluated. Looking back in the recent history of the values from light sensor 2302 , the readings were originally at level 2311 , which is below the light threshold {Threshold(light)}. However, more recently, the light level rose to the level at 2310 . As this level at 2310 is above the light threshold {Threshold(light)}, the values from the light sensor corroborate the event that may be have been detected by accelerometer 2301 . With respect to sound level, older sound level readings were at level 2315 , which were below the sound threshold {Threshold(sound)}. More recently, the sound level rose to level 2314 , which is above the sound threshold {Threshold(sound)}. Here, the output from the sound sensor also corroborates event that may have been detected by accelerometer 2301 .
With respect to both the light sensor 2302 and sound sensor 2303 , an individual signal value different from a maximum value above a threshold having been reached during a time interval is less relevant than the signal having reached the threshold during the time window. Stated differently, once it has been determined that a light signal is above the light threshold {Threshold(light)} for sensor reading 2310 , other readings between levels 2312 and 2313 are not considered for this threshold analysis. Similarly, variants between sound level 2316 and 2317 are less relevant than the sound level 2314 having passed the sound threshold level {Threshold(sound)} as the sound threshold has already been met.
Finally, FIG. 23 shows data dump points 2305 , 2306 , and 2307 after which insignificant signal readings are dumped from the memory of processor 100 and/or storage 105 . Interestingly, the data dump points 2305 , 2306 , and 2307 do not have to be at the same time window from the present. Rather each may have its own separate window length during which signal levels are maintained.
FIG. 24 shows an example of different techniques for monitoring core temperature including microwave radiometry and microwave thermometry in accordance with one or more aspects of the disclosure. For instance, core temperature 2401 may be determined through passive technologies including microwave radiometry 2402 in which energy from other sources is used to determine core temperature. Also, active techniques including microwave thermometry 2403 may be used to determine core temperature. For these two examples, separate antennas may be used for ultra-wideband device (UWB) and the microwave radiometry/thermography core temperature determination system as shown by state 2404 . Alternatively, a single antenna may be shared between the UWB and the core temperature determination device. For example, one or more switches may be used to alternatively connect the shared antenna to the UWB in the microwave radiometry/thermography core temperature determination system as shown by state 2405 .
UWB Modifications
One or more aspects of the disclosure relate to enhanced UWB operations to accommodate issues created by hair/fur, movement and mobility, air gaps, the curvature variations in necks of animals, and strap tightness (or closeness to the animal's skin).
For instance, the thickness and density of hair/fur, air gaps, and strap tightness pertain to a greater variance in the number of sets of ranges that may be used over conventional UWB systems (which generally require no air gap as direct skin contact is required). By increasing the number of discrete ranges used by the UWB system (for instance, by stepwise increasing the number of ranges (and possibly the overall range as well) the UWB radar may be modified to accommodate a large variability in spacing between the antennas and the observed tissue.
Next, to accommodate for the range increase different approaches may be used separately or in combination. For instance, the amplitude of the pulses may be increased to accommodate a greater need for power. Also, the pulse repetition frequency may also be increased until an acceptable signal to noise ratio is obtained. Further, these two approaches may be used in combination to provide a greater operation range of the UWB system while keeping the system compact and portable.
Next, to reduce unwanted emissions, the UWB may be triggered only when a number of other sensors/devices indicate that the firing of the radar is more likely than not to provide acceptable results. For instance, the UWB may not fire until the accelerometer indicates that the animal is moving below a given threshold (for instance, a threshold observed when the animal is sleeping). Also, the UWB may not fire until a thermometer on the unit indicates that a temperature facing the animal is above a threshold (for instance, a threshold being a temperature when the device is proximate the animal's neck while the animal is resting).
Next, heart rate variability in animals (including dogs) is higher than that of humans. For instance, a dog's heart rate may jump from 40 beats per minute to over 240 beats per minute in a short period (permitting explosive bursts of emery). To capture (or more accurately, to keep up with) this variability, the UWB system may include an adjustable window sampling size to monitor heartbeats. For instance, at 40 BMP, a window larger than 1.5 seconds per beat may accommodate that rate. However, at 240 BMP, the window needs to be closer to 0.25 seconds per beat. Accordingly, the system may include an auto-ranging window that is cycled through window sizes of 0.2 seconds through 2 seconds periodically, or even initially as the UWB is active.
Further, to account for different neck sizes, the UWB antenna may include a wide angle distribution pattern to accommodate the different sizes. Alternatively, different antennas may be used for different size necks. For instance, smaller animals may need the wider distribution antenna to accommodate a greater angle between different tissues being monitored while larger breeds may use a narrow field of view antenna that is more narrowly focused to a particular region. This may reduce interference from extraneous sources. Further, with additional antenna elements, the antenna may be steered toward different selective tissues for monitoring.
Further modifications may include the use of different radar generation procedures (for instance, using heterodyning processes) and/or the coding of pulses.
Owner's User Interface
FIGS. 25 and 26 show illustrative examples of an owner's user interface as displayable on a computer or smart phone. The Owner Health & Wellness Dashboard allows the owner to see in one place all trending information on the animal from sensor data and DMS derived data.
FIG. 25 shows a display 2501 of various information and conditions of a monitored animal in accordance with aspects of the disclosure. The display includes information drawn from both the wearable device 101 as well as from content from the veterinarian. For instance, information from the veterinarian includes the next scheduled appointment content 2502 and the identification of what medications are expiring next and the expiration dates. This information may help remind the user to keep the veterinarian appointment.
Next, the display 2501 includes content from the wearable device and/or the DMS in the form of instantaneous vital signs/physiological signs were overall trends relevant to the animal. For instance, display 2501 includes graphical indicators of activity 2505 , sleep 2506 , hydration 2507 , diet 2508 , stress 2509 , core temperature 2510 , weight 2511 , heart rate 2512 , and respiration rate 2513 . The following items relate to instantaneous vital signs/physiological signs from the wearable device: core temperature 2510 , heart rate 2512 , and respiration rate 2513 .
In contrast to the vital signs, the following items relate to wearable device-derived events or DMS-derived events such that they incorporate content from different sensors and may include tracking of health-related vital signs/physiological signs and/or activities over time: activity 2505 , sleep 2506 , hydration 2507 , diet 2508 , stress 2509 , and weight 2511 .
For purposes of illustration, each of the graphical displays of these items is shown as a dial with an arrow pivoting from one side of the dial to the other based on the state of the displayed item (e.g., a green area indicating no concern, a yellow area indicating caution, and a red area indicating concern for that individual item).
FIG. 26 shows activity level for that particular animal in accordance with aspects of the disclosure. The Owner Level Detail screen allows the owner to drill down on a specific item from the dashboard and review goals, alerts, recommendations, and more detailed, long term analyses information. For instance, the display 2601 of FIG. 26 includes an identification of the animal 2602 , a current indicator 2603 for the detail screen (in this example, the activity of the animal), and an alert message box 2604 identifying an alert determined by the wearable device 101 and or the DMS 301 (in this example that the animal missed two consecutive days of walks with an identification of the date and time of when the walks were missed). Next, the display 2601 may further include recommendations in field 2605 to improve the health of the animal (for instance, to resume daily walks). The display 2601 may include one or more goals as set by the veterinarian, the owner, or the DMS 301 . In this example, the goals are to walk 40 minutes per day, to keep the animal's weight below 80 pounds and to play 15 minutes. The display 2601 may further include an identification of the alert thresholds in field 2608 . In this example, the alert thresholds are missing two days of a walk, a change in gait dropping 15%, and an overall drop in activity of 25%.
Finally, a timeline of the displayed item of detail may be shown as content 2607 . Here, the timeline shows how the animal's activity level has changed over 12 weeks.
While the detailed screen 2601 of FIG. 26 relates to activity, it is appreciated that similar detail screens may be provided for other items identified in FIG. 25 with similar content including a graphical indication of the current status of that item, alerts, recommendations, goals, alert thresholds, and timelines.
Microwave Radiometry
A microwave radiometer is described below. As the available noise power even from a 50Ω termination on a 500 MHz bandwidth is only about −86 dBm, any significant RF loss in the signal path can induce significant errors. The architectures shown in FIG. 28 (specifically, FIGS. 28A and 28B ) were reviewed. Consistent changes were not observed in the output power when the phantom's temperature changed in the desired temperature range. See the inconsistent values of FIG. 29A . If the water temperature is increased (above 43° C.), the radiated power rises above the noise.
In order to increase the SNR in lower temperatures suitable to the collar-mounted environment, an amplifier located before the switch may be used. The architecture of the system is shown in FIG. 30 including a calibration antenna 3001 , amplifier 3002 , an active antenna 3007 and amplifier 3008 and switch 3009 . After the switch 3009 , an amplifier 3003 , a bandpass filter 3004 , power sensor/power meter 3005 , and RF processing system 3006 (the processing system 3006 as known in the art) follow. Output from the power sensor 3005 and/or processing system 3006 may be used individually or in combination via switch control 3010 to operate the switch 3009 . Switch 3009 toggles between top branch containing calibration antenna pointing towards the air and bottom branch containing antenna in contact with tissue/phantom under test.
An accurate microwave thermometer needs to detect low power levels of about −174 dBm/Hz at 37° C. (body temperature) emitting from the body. In our system, we selected the frequency range of 3.25-3.75 GHz which amounts to around −86 dBm power at the front-end and is far enough from common interference signals such as cellular and 2.4 GHz ISM band. The reflection coefficient of this antenna while in contact with air is under −10 dB in the 3.25-3.75 GHz bandwidth while it drops under −15 dB when it contacts or becomes proximate with either a body or a water phantom.
FIG. 31 shows a prototype of the microwave thermometer. The architecture of the system is shown in FIG. 31 including an antenna 3101 facing into open space (air), amplifier 3102 , a phantom (or water) antenna 3107 and amplifier 3108 and switch 3109 . After the switch 3109 , an amplifier 3103 , a series of filters 3104 (e.g., low pass 3104 a , high pass 3104 b , low pass 3104 c , high pass 3104 d ), power sensor/power meter 3105 , and RF processing system 3106 (the processing system 3106 as known in the art) follow. Other combinations of filters may be used to make a bandpass filter. Output from the power sensor 3105 and/or processing system 3106 may be used individually or in combination via switch control 3110 to operate the switch 3109 . Switch 3109 toggles between top branch containing calibration antenna pointing towards the air and bottom branch containing antenna in contact with tissue/phantom under test.
The subject under test may be a water phantom with the respective antenna in contact or in close proximity to it. Next is an amplifier for that antenna. For the other branch, the reference antenna is aimed at air (for instance, away from a subject or other body) and is followed by another amplifier. Next is a switch between the two antenna feeds, another amplifier, a cascade of filters making a bandpass filter (low pass, highpass, low pass and a final high pass filter). These are followed by the power sensor which is connected to the power meter. A final computing system is connected to the power meter. Not shown are the power supply (5V) and the ground wires connected to the active components, the switch control wire and the power supply.
The following RF components can be used in the system of FIG. 31 . There are three (ZX60-3800LN+) amplifiers and two low pass (VLF-3400+) and two high pass filters (VHF-3100+) from mini circuits. The four low-high-low-high pass filters in cascade may create a band pass filter with about −4 dB loss. Finally, the RF switch can be a TB-461+ switch also from mini circuits. Testing the RF front end results in a total gain of 41 dB in the bandwidth. The RF signal from filter's output may be connected to an Agilent E4412A power sensor, which is connected to an E4416A power meter. The power meter is connected to the computer. The power magnitude is gathered and recorded by the computer.
The system is tested as follows: first, the RF switch is toggled to route the signal from the calibration antenna (pointing to the air) to the power meter. The output power is detected in (for instance) a 10 second period and averaged creating a value called Pcalib, which will be used for offset adjustment.
The switch is then toggled to the contact antenna in contact with the phantom. In this test, the phantom is a 1% NaCl water solution in a glass beaker. The antenna is strapped around the beaker. The water is heated to a desired temperature and the data is gathered in a 10 second period and averaged (Pout). The reference is a HH12B digital thermometer from Omega Engineering, which probe is inserted into the water solution and records the temperature of the water. The temperature of each of the amplifiers is kept constant.
For testing purposes in one example, the measurements are performed in 5 different days. The data of the first three days are used to build a linear model and then this model is tested on all 5 days of data.




s
=



-
0.1748937







P
calib


-
6.6062371







b
=


6.81888104






P
calib


+
217.721194








T
1

=


(


P
out

-
b

)

s







{






T
w

=

T
1


,


if






T
1


>
39









T
w

=


(


T
1

-
25.80058962

)

0.351641


,


if






T
1


<
39









The measured temperature (Tw) versus the actual temperature ° F. is shown in FIG. 32A and the error has been shown in FIG. 32B . The result is that the measured temperature using this structure has under 2° F. error in 61% of the time.
Measurements show that the change in RF power in a 9° C. temperature range is about 0.35 dBm. The power meter used in the test bench is conventional with an accuracy of 0.0001 dBm. An alternative, off-the-shelf (1.75″×1.25″ size) power meter may include a ZX47-60LN+ from mini-circuits with the detection graph in FIG. 33 . In order to get around 0.025 dBm accuracy from this hardware (assuming no performance degradation in RF performance after integration of hardware components on a printed circuit board (PCB)), a 12-bit analog to digital converter (with 2.5 V input voltage range) may be used. This accuracy is included in the ADCs of the microprocessor currently used in the board. Based on this information, this radiometer may be implemented with the 2° F. accuracy in 61% of the time while with repeatable measurements the non-accurate results can be detected and discarded.
All components up to the detector and the antenna may be fitted on 1.7″×3.5″ PCB. Additional space for the power meter may increase the size of the PCB to approximately 2″×3.7″. FIG. 35 shows readings taken for a core temperature and a reference over time.
FIGS. 36A , 36 B, and 36 C show three sets of readings for core temperature and a reference over time. A bias (offset) is present when using real time calibration.
FIGS. 37A and 37B show graphs of real temperature compared to measured temperature and the error compared to water temperature, respectively. Readings of lower temperatures appear to have a greater error than readings of higher temperatures.
FIG. 38 shows measurements over time of two subjects (two dogs) and the environment for each using a thermocouple.
Alternative approaches to improve accuracy include: using different models for universal calibration; and correlating calibration data to increase accuracy of a model.
Further approaches include: increasing the switching frequency between the reference and the subject antenna; using other antenna designs; changing the reference to something other than air; modifying the components and/or architecture; adding a circulator (however, additional loss may be destructive); and removing one or more of the amplifiers before the switch (this may address some of the bias).
For reference, black body radiation in the region of room temperature (70° F., 21° C., or 343° K) is primarily infrared and is limited to near-surface temperatures while microwave can provide thermal information to several centimeters.
Black body radiation in the microwave spectrum in the region of room temperature is bandwidth dependent and defined by:
P dBm =30+10 log 10 ( k B T )+10 log 10 (Δ f )=−174 dBm+10 log 10 (Δ f )












F (Hz)
P (dBm)











1
−173.83


10
−163.83


100
−153.83


1000
−143.83


1.00E+04
−133.83


1.00E+05
−123.83


1.00E+06
−113.83


1.00E+07
−103.83


1.00E+08
−93.83


1.00E+09
−83.83






































Black Body Temperature
Power (dBm) versus Bandwidth












° F.
° C.
° K.
1 Hz
100 Hz
10 kHz
1 MHz
100 MHz
500 MHz
pW














90
32.2
305.9
−173.74
−153.74
−133.74
−113.74
−93.74
−86.753


91
32.8
306.5
−173.73
−153.73
−133.73
−113.73
−93.73
−86.745


92
33.3
307.0
−173.73
−153.73
−133.73
−113.73
−93.73
−86.737


93
33.9
307.6
−173.72
−153.72
−133.72
−113.72
−93.72
−86.729


94
34.4
308.1
−173.71
−153.71
−133.71
−113.71
−93.71
−86.721


95
35.0
308.7
−173.70
−153.70
−133.70
−113.70
−93.70
−86.713
2.1314

96
35.6
309.2
−173.70
−153.70
−133.70
−113.70
−93.70
−86.705
2.1353

97
36.1
309.8
−173.69
−153.69
−133.69
−113.69
−93.69
−86.698
2.1391

98
36.7
310.3
−173.68
−153.68
−133.68
−113.68
−93.68
−86.690
2.1429

99
37.2
310.9
−173.67
−153.67
−133.67
−113.67
−93.67
−86.682
2.1468

100
37.8
311.5
−173.66
−153.66
−133.66
−113.66
−93.66
−86.674
2.1506

101
38.3
312.0
−173.66
−153.66
−133.66
−113.66
−93.66
−86.667
2.1545

102
38.9
312.6
−173.65
−153.65
−133.65
−113.65
−93.65
−86.659
2.1583

103
39.4
313.1
−173.64
−153.64
−133.64
−113.64
−93.64
−86.651
2.1621

104
40.0
313.7
−173.63
−153.63
−133.63
−113.63
−93.63
−86.643
2.1660

105
40.6
314.2
−173.63
−153.63
−133.63
−113.63
−93.63
−86.636
2.1698

106
41.1
314.8
−173.62
−153.62
−133.62
−113.62
−93.62
−86.628
2.1736

107
41.7
315.3
−173.61
−153.61
−133.61
−113.61
−93.61
−86.620
0.042fW

108
42.2
315.9
−173.60
−153.60
−133.60
−113.60
−93.60
−86.613


109
42.8
316.5
−173.59
−153.59
−133.59
−113.59
−93.59
−86.605


110
43.3
317.0
−173.59
−153.59
−133.59
−113.59
−93.59
−86.598










0.085
dBm










































For instance, desired accuracy for consumers may be within 1° F. for consumers and 0.5° F. for veterinarians. Accuracy may be improved by one or more of the following:
A. Stringent control/characterization of receiver bandwidth
B. Better antenna focus (i.e. narrow beamwidth)
C. Better coupling of antenna to test subject
D. Consistent coupling of antenna to test subject
E. Regular and frequent calibration of system
F. Minimization of circuit drift—voltages, temperatures, etc.
G. Highly deterministic system
H. Accurate models of test subject and antenna
Although example embodiments are described above, the various features and steps may be combined, divided, omitted, and/or augmented in any desired manner, depending on the specific secure process desired. This patent should not be limited to the example embodiments described, but rather should have its scope determined by the claims that follow.