Search

RU-2861611-C1 - METHOD FOR MONITORING ANIMAL HEALTH

RU2861611C1RU 2861611 C1RU2861611 C1RU 2861611C1RU-2861611-C1

Abstract

FIELD: veterinary medicine. SUBSTANCE: group of inventions relates to monitoring animal health. Load data may be obtained from a plurality of load sensors associated with a platform holding litter contained thereon, wherein individual load sensors from the plurality of load sensors are spaced apart from each other and receive pressure input from the platform independently of each other. If it is determined that the load data is obtained as a result of or not as a result of animal interaction with the contained litter, then an animal behaviour characteristic associated with the animal is recognised if, based on the load data, it is determined that the interaction with the contained litter is due to animal interaction. The animal behaviour characteristic is classified into a classified animal event using a machine learning classifier. A change in the classified animal event is identified compared to a previously recorded event associated with the animal. EFFECT: increasing the accuracy of animal health monitoring results. 33 cl, 12 dwg

Inventors

  • DONAVON, MARK ALAN
  • NASANURU, Abhishek Sai
  • KRISHNAN, Ayushi
  • RAVI, Dwarakanath Raghavendra
  • SHERWOOD, Daniel James
  • MAGUIRE, Russell Stewart
  • STONE, Jack William James
  • LOGAN, Georgina Elizabeth Mary
  • HATORI, Tomoko
  • HAUBRICK, Peter Michael
  • SCHIM VAN DER LOEFF, Wendela Sophie
  • LANGENFELD-MCCOY, Natalie
  • MCGOWAN, Ragen Trudelle-Schwarz
  • DUSSAN, Helber
  • KAMARAJ, Mani Bharath
  • VIJAYARAJAN, Vignesh
  • GOVINDARAJAN, Venkatakrishnan
  • SINGH, AJAY
  • MALIPEDDI, Sarath

Dates

Publication Date
20260506
Application Date
20220826
Priority Date
20210827

Claims (20)

  1. 1. A method of monitoring the health of an animal, controlled by at least one processor, comprising: receiving load data from a plurality of load sensors of an animal monitoring device, wherein the plurality of load sensors are coupled to a platform supporting litter contained thereon, wherein individual load sensors of the plurality of load sensors are separate from one another and receive pressure input data from the platform independently of one another; determining whether the load data is obtained as a result of an animal interacting with the contained litter; determining a location of the animal within a litter box disposed on the platform, wherein the location of the animal within the litter box is determined based on the location of the center of gravity of the animal within the litter box at various times during the animal interacting; tracking the center of gravity of the animal; recognizing a behavioral characteristic of the animal associated with the animal if it is determined based on the load data that the interaction with the contained litter is due to an animal interacting; classifying the behavioral characteristic of the animal into a classified animal event using a machine learning classifier; and identifying a change in the classified animal event compared to a previously recorded event associated with the animal; wherein classifying the animal behavior characteristic further comprises analyzing load data obtained from the plurality of load sensors to determine a pattern of movement of the animal during the corresponding classified animal event, wherein the pattern of movement is determined based on the tracked movement of the center of gravity of the animal; and identifying the animal based on the determined pattern of movement.
  2. 2. The method of claim 1, wherein the classification of the animal's behavior includes a litter box event, a urination event, a defecation event, a non-defecation event, or a combination thereof.
  3. 3. The method of claim 1, further comprising correlating a change in the classified event of the animal with a physical, behavioral, or mental disorder associated with the animal.
  4. 4. The method of claim 3, wherein the physical disorder is a disease of the animal.
  5. 5. The method of claim 4, wherein the animal's disease is a feline disease selected from urinary tract disease, kidney disease, diabetes, hyperthyroidism, idiopathic cystitis, digestive problems, or arthritis.
  6. 6. The method according to claim 3, wherein the mental disorder is selected from anxiety, stress, or cognitive impairment.
  7. 7. The method of claim 3, wherein the behavioral disorder is performing natural bowel movements outside the litter box.
  8. 8. The method of claim 1, wherein determining whether the load data is the result of an animal interaction determines whether the load data is the result of an animal interaction, a human interaction, a false alarm, or a random interaction.
  9. 9. The method of claim 1, wherein the identification of the animal distinguishes the animal from at least one other animal that interacts with the platform.
  10. 10. The method of claim 1, further comprising generating a notification indicating a change in the classified animal event.
  11. 11. The method of claim 10, wherein the notification is generated after a parameter associated with a device event reaches a threshold value.
  12. 12. The method of claim 1, which does not include or communicate with any camera or image capture device and does not perform visual image recognition.
  13. 13. The method of claim 1, wherein classifying the animal behavior characteristic further comprises analyzing load data from a plurality of load sensors to measure the mass of the litter box located on the platform, the mass distribution of the animal, the location of the event, the duration of the event, the movement pattern, the entry force, the exit force, the instability of the animal interaction, or a combination thereof.
  14. 14. The method of claim 1, wherein classifying the animal behavior characteristic further comprises analyzing load data from a plurality of load cells to identify or measure entry of the animal into the litter box on the platform, the amount of animal movement to select a particular location to eliminate, the amount of time to select a particular location to eliminate, the amount of time spent preparing a particular location to eliminate, the amount of energy spent preparing a particular location to eliminate, the amount of time spent covering the elimination, the amount of energy spent covering the elimination, the duration of the elimination, the total duration of the device event from entry to exit of the animal, the mass of the elimination, the movement of the animal during the elimination, the detection of a step/tilt on one load cell during the elimination, the exit of the animal from the installed litter box, (xii) one or more movements or impacts affecting the litter box, or a combination thereof.
  15. 15. The method of claim 1, wherein classifying the animal behavior characteristic further comprises analyzing load data from a plurality of load sensors in both the time domain based on a time domain feature and the frequency domain based on a frequency domain feature.
  16. 16. The method of claim 15, wherein the time domain feature comprises a mean, a median, a standard deviation, a range, an autocorrelation, or a combination thereof, and wherein the time domain feature is generated as input information or input data for a machine learning classifier.
  17. 17. The method of claim 15, wherein the frequency domain feature comprises a median, an energy, a power spectral density, or a combination thereof, and wherein the frequency domain feature is generated as input information or input data for a machine learning classifier.
  18. 18. The method of claim 15, wherein the selected time domain features and the frequency domain features are selected from a standard deviation of the load, a plateau length, a number of mean crossings, a unique number of peaks, a number of distinct load values, a ratio of distinct load values to event duration, a number of maximum load changes in individual sensors, a percentage of an average load interval, a percentage of a high load interval, a high load interval instability, a high load interval offset, an autocorrelation function lag or delay, a curvature, a linearity, (xv) a number of peaks, an energy, a minimum power, a standard deviation of the power, a maximum power, a largest offset of the offset, a maximum Kullback-Leibler divergence, a Kullback-Leibler divergence time, a spectral density entropy, an autocorrelation function differentials, an autoregressive model change, or a combination thereof; and classifying the interaction of the animal, determining the identity of the animal, or both based on using the selected time domain features and the selected frequency domain features as input information or input data to a machine learning classifier.
  19. 19. The method of claim 1, wherein classifying the animal behavior characteristic further comprises analyzing load data from a plurality of load sensors at the overall load, individual loads per load sensor, at the phase level by means of a phase separation algorithm that separates the load data into phases, or a combination thereof.
  20. 20. The method of claim 19, wherein the phase separation algorithm that separates the load data into phases includes at least three phases, including preparation for the execution of natural dispatches, the execution of natural dispatches, and after the execution of natural dispatches.

Description

This application claims priority to U.S. Provisional Patent Application No. 63/237,664, filed August 27, 2021, which is incorporated herein by reference in its entirety. PREREQUISITES FOR THE CREATION OF THE INVENTION Cats typically use litter boxes for natural elimination—urination and defecation. The litter box contains a layer of cat litter, which collects urine and feces. Pet litter contains absorbent and/or adsorbent material, which can be clumping or non-clumping. Visual indicators associated with litter box use can provide information about a cat's health, such as the development of physical, behavioral, or mental disorders. Unfortunately, these symptoms may only appear in the middle or late stages of illness or health problems and often do not provide sufficient information for appropriate intervention. Furthermore, pet owners often lack the knowledge about animal behavior to associate litter box use with health issues. Some attempts have been made to track litter box activity as a means of assessing cat health. For example, cameras, video recorders, and/or scales have been used to record litter box activity. While these devices can be useful in tracking some basic information about cat behavior, they typically provide one-dimensional information, may require a trained behaviorist to interpret, and/or may not provide reliable data on less obvious and/or invisible cues. BRIEF DESCRIPTION OF GRAPHIC MATERIALS Figs. 1A-1C are schematic representations of examples of animal health monitoring systems in accordance with the present disclosure. Fig. 2 provides a general overview of examples of events that may occur using animal health monitoring systems in accordance with the present description. Figs. 3A–3E show examples of load signals for cat-in-litter events in accordance with the present description. Figs. 4A–4C show examples of load signals for cat-out-of-litter events in accordance with the present disclosure. Figs. 5A-5B show examples of load signals for waste removal events in accordance with the present description. Figs. 6A–6B show examples of load signals for movement events in accordance with the present description. Fig. 7 shows examples of phases within an event in accordance with the present description. Fig. 8 shows an example of a flow chart of a method for classifying animal behavior in accordance with the present description. Fig. 9A shows a location on the trajectory of the animal's movement in accordance with the present description. Figs. 9B-9C illustrate the identification of animals based on animal behavior in accordance with the present description. Fig. 10 shows a block diagram of the method for identifying an animal in accordance with the present description. Fig. 11 shows the characteristics of various classification models in accordance with the present description. Fig. 12 is a block diagram of a method for monitoring the health of an animal in accordance with the present description. DETAILED DESCRIPTION The present description relates to the field of animal health and behavior monitoring and, in particular, to devices, systems, methods and computer software products for determining, monitoring, processing, recording and transmitting over a network various physiological and behavioral parameters of animals. According to the examples of the present disclosure, a method for monitoring the health of an animal under the control of at least one processor is described. The method may include receiving load data from a plurality of load sensors associated with a platform supporting litter contained thereon. Individual load sensors of the plurality of load sensors may be separate from one another and may receive pressure input data from the platform independently of one another. The method may further include determining whether the load data was obtained as a result of an animal's interaction with the contained litter. The method may further include recognizing a behavioral characteristic of the animal associated with the animal if, based on the load data, it is determined that the interaction with the contained litter is due to an animal's interaction. The method may further include classifying the behavioral characteristic of the animal into a classified animal event using a machine learning classifier. The method may further include identifying a change in the classified animal event compared to a previously recorded event associated with the animal. In another example, the present description provides a non-transitory computer-readable storage medium having instructions embodied thereon, which, when executed, cause a processor to perform a method for monitoring the health of an animal. The method may include receiving load data from a plurality of load sensors associated with a platform supporting litter contained thereon, wherein individual load sensors of the plurality of load sensors are separate from one another and receive input pressure data independently of one another.