System and method for animal health monitoring

Abstract

The present disclosure provides systems and methods for animal health monitoring. Load data can be obtained from an animal monitoring device comprising three or more load sensors associated with a platform on which contained litter is carried, wherein each of the three or more load sensors are separate from one another and receive pressure input from the platform independently of one another, wherein the three or more load sensors individually sample load at 2.5 Hz to 110 Hz. If it is determined based on the load data that the interaction with the contained litter is due to an animal interacting with the contained litter, an animal behavioral attribute related to the animal can be identified. The animal behavioral attribute can be classified as an animal classification event using a machine learning classifier.

Description

[0001] This application claims the benefit of U.S. Provisional Patent Application No. 63 / 237,664, filed August 27, 2021, the entire contents of which are incorporated herein by reference. Background Technology

[0002] Cats use litter boxes to eliminate urine and feces. A litter box consists of a layer of cat-specific bedding to receive urine and feces. Pet bedding contains absorbent and / or absorbent materials, which can be non-clumping or clumping. Visual indicators associated with litter box use can provide information about a cat's health status; for example, the onset of physical, behavioral, or mental health problems. Unfortunately, these symptoms may only appear in the middle or later stages of an illness or health problem, often not providing sufficient information for proper intervention. Furthermore, pet owners often lack knowledge of animal behavior and cannot link litter box use to health problems.

[0003] Some efforts have been made to track litter box activity to assess a cat's health. For example, cameras, video recording devices, and / or scales have been used to capture litter box activity. While these devices may help track some basic information about a cat's behavior, they typically provide one-dimensional information that may require interpretation by a qualified behaviorist and / or may not provide good data on more subtle and / or non-visual cues. Attached Figure Description

[0004] Figures 1A to 1E An exemplary animal health monitoring system according to this disclosure is illustrated schematically.

[0005] Figure 2 A conceptual overview of exemplary events that may occur when using an animal health monitoring system according to this disclosure is shown.

[0006] Figures 3A to 3E An exemplary load signal for a cat in a box event according to this disclosure is shown.

[0007] Figures 4A to 4C An exemplary load signal for a cat outside the box according to this disclosure is shown.

[0008] Figures 5A to 5B An exemplary load signal for a removal event according to this disclosure is shown.

[0009] Figures 6A to 6B An exemplary load signal for a movement event according to this disclosure is shown.

[0010] Figure 7 Exemplary phases within an event according to this disclosure are shown.

[0011] Figure 8An exemplary flowchart of a method for classifying animal behavior according to this disclosure is shown.

[0012] Figure 9A Exemplary locations of animal movement paths according to this disclosure are shown.

[0013] Figures 9B to 9C Examples of animal identification based on animal behavior according to this disclosure are shown.

[0014] Figure 10 A flowchart illustrating an example of animal identification according to this disclosure is shown.

[0015] Figure 11 Examples of the performance of various classification models according to this disclosure are shown.

[0016] Figures 12 to 14 A flowchart of an exemplary method for monitoring the health status of an animal according to this disclosure is shown. Detailed Implementation

[0017] This disclosure relates to the field of animal health and behavior monitoring, and more specifically to apparatus, systems, methods, and computer program products for determining, monitoring, processing, recording, and transmitting various physiological and behavioral parameters of animals via networks.

[0018] According to examples of this disclosure, a method for monitoring the health status of an animal under the control of at least one processor is disclosed. The method may include obtaining load data from an animal monitoring device including three or more load sensors associated with a platform on which bedding material is carried, wherein each of the three or more load sensors is separate from and independently receives pressure input from the platform, and wherein the three or more load sensors individually sample the load at a frequency of 2.5 Hz to 110 Hz. The method may further include identifying animal-related animal behavioral attributes if it is determined from the load data that the interaction with the bedding material is due to the interaction between the animal and the bedding material. The method may further include classifying the animal behavioral attributes into animal classification events using a machine learning classifier.

[0019] In another example, this disclosure provides an animal monitoring system including an animal monitoring device. The animal monitoring device may include a platform configured to carry bedding material thereon. The device may also include three or more load sensors associated with the platform, wherein each of the three or more load sensors is separate from and independently receives pressure input from the platform, and wherein each of the three or more load sensors individually has a sampling rate in the range of 2.5 Hz to 110 Hz. The device may also include a data communicator configured to independently transmit load data from the three or more load sensors.

[0020] In another example, this disclosure provides a method for monitoring the health status of an animal under the control of at least one processor. The method may include obtaining load data from an animal monitoring device including three or more load sensors associated with a platform on which contained bedding material is held, wherein each of the three or more load sensors is separate from and independently receives pressure input from the platform. The method may also include identifying animal-related animal behavioral attributes if it is determined from the load data that the interaction with the contained bedding material is due to the interaction between the animal and the contained bedding material. The method may further include classifying the animal behavioral attributes into animal classification events using a machine learning classifier, including analyzing the load data via a stage separation algorithm, wherein animal classification events include animal excretion, and wherein the stage separation algorithm is capable of classifying multiple discrete animal excretions occurring at multiple discrete animal excretions.

[0021] In another example, this disclosure provides an animal monitoring system including an animal monitoring device. The animal monitoring device may include a platform configured to carry bedding material thereon. The animal monitoring device may also include three or more load sensors associated with the platform, wherein each of the three or more load sensors is separate from and independently receives pressure input from the platform. The animal monitoring device may also include a data communicator configured to independently transmit load data from the three or more load sensors. The animal monitoring device may also include a processor and a memory storing instructions, which, when executed by the processor, include acquiring load data independently of the three or more load sensors. The instructions also include identifying an animal-related animal behavioral attribute if it is determined from the load data that the interaction with the platform is due to animal interaction. The instructions also include classifying the animal behavioral attribute into animal classification events using a machine learning classifier, including analyzing the load data via a stage separation algorithm, wherein animal classification events include animal excretion, and wherein the stage separation algorithm is capable of classifying multiple discrete animal excretions occurring at multiple discrete animal excretions.

[0022] In another example, this disclosure provides a non-transitory machine-readable storage medium having instructions thereon that, when executed, cause a processor to perform a method for monitoring the health status of an animal. The method may include obtaining load data from an animal monitoring device including three or more load sensors associated with a platform on which bedding material is carried, wherein each of the three or more load sensors is separate from and independently receives pressure input from the platform. The method may also include identifying animal-related animal behavioral attributes if it is determined from the load data that the interaction with the bedding material is due to an interaction between the animal and the bedding material. The method may further include classifying the animal behavioral attributes into animal classification events using a machine learning classifier, including analyzing the load data via a stage separation algorithm, wherein animal classification events include animal excretion, and wherein the stage separation algorithm is capable of classifying multiple discrete animal excretions occurring at multiple discrete animal excretions.

[0023] Additional features and advantages of the disclosed methods and apparatus are described herein, and these features and advantages will become apparent from the following detailed description and accompanying drawings. Not all features and advantages described herein are included; in particular, many additional features and advantages will be apparent to those skilled in the art from the accompanying drawings and description. Furthermore, it should be noted that the language used in this specification has been chosen primarily for readability and instructional purposes and does not limit the scope of the subject matter of the invention.

[0024] definition

[0025] As used herein, “about,” “approximately,” and “substantially” should be understood to refer to a number within a range of values, such as -10% to +10%, -5% to +5%, -1% to +1%, or -0.1% to +0.1% of the mentioned number. All numerical ranges herein should be understood to include all integers, non-negative integers, or fractions within that range. Furthermore, these numerical ranges should be understood to support claims that involve any number or subset of numbers within that range. For example, disclosures of 1 to 10 should be understood to support ranges of 1 to 8, 3 to 7, 1 to 9, 3.6 to 4.6, 3.5 to 9.9, etc.

[0026] As used in this disclosure and the appended claims, the singular forms “a,” “an,” and “the” include plural references unless the context clearly specifies otherwise. Thus, for example, references to “a component” or “the component” include two or more components.

[0027] The term "including / comprises" will be interpreted as inclusive rather than exclusive. Similarly, the terms "including / comprises" and "or" should be considered inclusive unless the context explicitly prohibits this interpretation. Therefore, the disclosure of embodiments using the term "including / comprises" includes both embodiments that "consist substantially of the specified components" and embodiments that "consist of the specified components".

[0028] The term “and / or” used in the context of “X and / or Y” should be interpreted as “X” or “Y” or “X and Y”. Similarly, “at least one of X or Y” should be interpreted as “X” or “Y” or “X and Y”.

[0029] In the context of this document, the terms “example” and “such as” (especially when followed by a list of terms) are exemplary and illustrative only and should not be considered exclusive or comprehensive.

[0030] The terms “pet” and “animal” are used synonymously herein and refer to any animal that may use a litter box, including, but not limited to, cats, dogs, mice, ferrets, hamsters, rabbits, lizards, pigs, or birds. A pet can be any suitable animal, and this disclosure is not limited to any particular pet animal. The term “excrement” refers to urine and / or feces excreted by a pet.

[0031] As used herein, the term "bedding" refers to any substance that absorbs animal urine and / or reduces the odor from animal urine and / or feces. "Clumping bedding" forms aggregates in the presence of moisture, wherein these aggregates are distinct from the other bedding in the bedding box. "Clumping agents" bind with adjacent particles when wet. "Non-clumping bedding" does not form distinct aggregates.

[0032] The term "bedding box" refers to any device that can hold pet bedding, such as a container having a bottom wall and one or more side walls, and / or any device constructed for placing bedding, such as a mat or grid, thereon. As a non-limiting example, a bedding box can be a rectangular box with side walls having a height of at least about six inches.

[0033] Animal health monitoring

[0034] According to this disclosure, systems and methods for monitoring animal health can be based on the location where animals typically defecate. For example, an animal health monitoring system for cats can typically be placed under the cat's litter box. This can be particularly beneficial because this configuration allows pet owners to use their existing litter boxes and cat litter, thereby minimizing any risk of cat defecation behavior problems that may occur when changing litter boxes. However, in other examples, the system and methods can also be performed using a new litter box or even a litter box integrated with or designed / formed for use with the platform and load sensors of this disclosure. In further detail, although the systems and techniques described herein are directed to cats and cat behavior, it should be noted that the systems and techniques described herein can be used to monitor the behavior of any animal.

[0035] In the examples disclosed herein, the animal health monitoring system may include one or more load sensors. The load sensors can monitor the weight distribution of animals within the animal health monitoring system and the time animals spend within the area monitored by the system. For example, load sensor data can be used to track cat movement patterns in litter boxes, identify non-cat interactions with the box, identify individual cats in a multi-cat scenario, identify litter box maintenance events, and / or predict multiple unique insights for each cat / litter box event. Based on this information, multiple events describing animal behavior can be identified. For example, it can be determined whether the load sensor data originates from cat behavior and / or human interaction with the litter box. If the behavior is cat-related, it can be determined whether the cat is interacting with the inside or outside of the litter box. If the cat is inside the litter box, the cat's identity and / or activity (urination, defecation, etc.) can be determined. If the cat is outside the litter box, multiple behaviors (e.g., rubbing the box, maintaining balance on the edge of the box, etc.) can be identified. If the behavior is human-related, it can be determined whether the person is shoveling litter, adding litter, interacting with the litter box, interacting with the animal health monitoring system, etc.

[0036] Animal health monitoring systems can automatically track visit frequency, visit type (e.g., excretion vs. non-excretion), and / or animal weight over multiple visits. This historical information can be used to monitor animal weight, litter box visit frequency, and / or excretion behavior over time. This information, optionally combined with various other data about the animal (e.g., age / life stage, sex, reproductive status, physical condition, rate of change in weight or behavior, etc.), can be used to identify when changes occur and / or predict potential health or behavioral conditions affecting the animal.

[0037] In addition to identifying animal behavior, animal health monitoring systems can advantageously provide early indicators of underlying health conditions, including but not limited to the animal's physical, behavioral, and mental health. Examples of physical health include, but are not limited to, kidney health, urinary tract health, metabolic health, and digestive health. More specifically, animal diseases that may be associated with weight and behavioral data obtained from the use of animal health monitoring systems include, but are not limited to, feline lower urinary tract diseases, diabetes, irritable bowel syndrome, feline spontaneous cystitis, bladder stones, bladder crystals, arthritis, hyperthyroidism, diabetes, and / or a variety of other diseases that may potentially affect the animal. Examples of behavioral health include, but are not limited to, out-of-cradle excretion and / or feline social dynamics in multi-cat households. Examples of mental health include, but are not limited to, anxiety, stress, and cognitive decline. Based on these underlying health conditions, proactive notification can be provided to the animal's owner and / or veterinarian for further diagnosis and treatment.

[0038] The animal health monitoring systems and technologies described herein offer several advantages over existing systems (although it is noted that the systems and methods described herein can be used in combination with some of these existing monitoring systems in certain situations). Existing monitoring systems typically rely on microchips implanted in the animal, RFID-enabled collars, and / or visual image recognition to identify individual cats. These systems can be highly invasive (e.g., requiring veterinary intervention to implant a microchip in a specific location within the animal), prone to failure (e.g., the microchip may migrate to another location within the animal and become difficult to locate; RFID collars may wear out, be lost, and / or require frequent battery replacements / recharging; cameras may require precise positioning and maintenance, etc.), and / or be highly disruptive to typical animal behavior. For example, the presence of camera systems or human observers and / or audible noise can prevent some cats from using their litter boxes in ways they would normally prefer. Furthermore, some existing systems require the use of specific materials (such as specific types of litter).

[0039] The animal health monitoring system disclosed herein addresses some limitations of existing systems, particularly where some of these other systems interfere with the normal behavior of animals. The animal health monitoring system of this disclosure can, for example, identify and track animals without relying on external identification, such as microchips or RFID collars. Furthermore, in some examples, the animal health monitoring system described herein can identify animals and their behavior without relying on image or video information, thereby avoiding the use of cameras or human observers that can influence the typical behavior of animals. For example, the animal health monitoring system provided herein can identify individual animals from a group of animals. In other words, the animal health monitoring system can distinguish each cat in a multi-cat household and provide it with independent health monitoring. In several embodiments, the animal health monitoring system includes more than one load sensor, thereby allowing the generation of more detailed information about the animals and their movement patterns compared to existing systems. For illustration, the sensors used in the animal health monitoring system are located in positions that do not disrupt the cat's natural behavior. The animal health monitoring system is designed with a low profile to accommodate even very young or old cats, as these cats may have difficulty entering boxes with a higher profile. Furthermore, the animal health monitoring system can utilize the cat's existing litter box and can be used with any type of litter (e.g., clumped or non-clumped), thus avoiding potential elimination problems when changing litter types. The system can be battery-powered or powered by a mains outlet, allowing use in locations without electrical outlets, eliminating tripping hazard power cords, or allowing cats prone to chewing on electrical cords to use it.

[0040] Now turn to the attached image. Figure 1A An animal health monitoring system 100 is schematically illustrated. The animal health monitoring system may include a client device 110 communicating via a network 140, an analysis server system 120, and / or an animal monitoring device 100. In this example, a litter box or container 132 containing litter 134 is placed on top of the animal monitoring device. The litter may be cat litter. In some examples, the analysis server system may be implemented using a single server. In others, the analysis server system may be implemented using multiple servers. In still other examples, the client device may interact with and utilize the analysis server system, and vice versa.

[0041] Client device 110 may include, for example, a desktop computer, laptop computer, smartphone, tablet computer, and / or any other user interface suitable for communicating with animal monitoring devices. The client device may obtain various data from one or more animal monitoring devices 130, provide data and insights about one or more animals via one or more software applications, and / or provide data and / or insights to the analytics server system 120, as described herein. The software applications may provide data on animal weight and behavior, track changes in data over time, and / or provide predictive health information about the animals, as described herein. In some embodiments, the software applications obtain data from the analytics server system for processing and / or display.

[0042] The analytics server system 120 can acquire data from various client devices 110 and / or animal monitoring devices 130, as described herein. The analytics server system can provide data and insights about one or more animals, and / or transmit data and / or insights to client devices, as described herein. These insights may include, but are not limited to, insights about animal weight and behavior, changes in data over time, and / or predictive health information about the animals, as described herein. In several embodiments, the analytics server system acquires data from multiple client devices and / or animal monitoring devices, identifies animal cohorts within the acquired data based on one or more characteristics of the animals, and determines insights into the animal cohorts. Insights into the animal cohorts can be used to provide recommendations for specific animals having characteristics common to the characteristics of the cohort. In many embodiments, the analytics server system provides veterinarians with a portal (e.g., a website) to access information about specific animals.

[0043] Animal monitoring device 130 can acquire data about animal and / or human interactions with the animal monitoring device. In some embodiments, the animal monitoring device includes a waste disposal area (e.g., a litter box) and one or more load sensors. In several embodiments, the load sensors include motion detection devices, accelerometers, weight detection devices, etc. The load sensors can be located in a position that does not disrupt the cat's natural behavior. The load sensors can automatically detect the presence of a cat in the litter box, and / or automatically measure the cat's characteristics when the cat is in the litter box or after the cat leaves the litter box. Additionally, the load sensors can be positioned to track the movement of the animal within the litter box. Data captured using the load sensors can be used to determine animal defecation behavior, behaviors other than defecation that may occur inside or outside the litter box (e.g., the cat rubbing against the litter box), and / or other environmental activities as described herein. The animal monitoring device can transmit data to client device 110 and / or analysis server system 120 for processing and / or analysis. In some examples, the animal monitoring device can communicate directly with a non-networked client device 115 without sending data over network 140. The term "non-networked" client device does not mean that it is not connected to the cloud or other networks, but simply that there is a wireless or wired connection that can directly connect to the animal monitoring device. For example, the animal monitoring device and the non-networked client device can communicate via Bluetooth. In some implementations, the animal monitoring device processes load sensor data directly. In many implementations, the animal monitoring device uses load sensor data to determine if the animal monitoring device is unbalanced. In this case, automatic or manual adjustment of one or more adjustable feet can rebalance the animal monitoring device. In this way, the animal monitoring device can adjust its own positioning to provide a stable platform for the waste disposal area.

[0044] Figure 1A Any computing device shown (e.g., client device 110, analysis server system 120, and animal monitoring device 130) may include a single computing device, multiple computing devices, a cluster of computing devices, etc. A computing device may include one or more physical processors communicatively coupled to memory devices, input / output devices, etc. As used herein, a processor may also be referred to as a central processing unit (CPU). The client device may be accessed by an animal owner, veterinarian, or any other user.

[0045] Additionally, as used herein, a processor may include one or more means capable of executing instructions that encode arithmetic, logic, and / or I / O operations. In one exemplary example, the processor may implement a von Neumann architecture model and may include an arithmetic logic unit (ALU), a control unit, and multiple registers. In many aspects, the processor may be a single-core processor typically capable of executing one instruction at a time (or a single pipeline of instructions) and / or a multi-core processor capable of executing multiple instructions simultaneously. In some examples, the processor may be implemented as a single integrated circuit, two or more integrated circuits, and / or may be a component of a multi-chip module, wherein the individual microprocessor dies are included in a single integrated circuit package and thus share a single socket. As discussed herein, memory refers to volatile or non-volatile memory devices, such as RAM, ROM, EEPROM, or any other means capable of storing data. Input / output devices may include network devices (e.g., network adapters or any other components that connect a computer to a computer network), peripheral component interconnect (PCI) devices, storage devices, disk drives, audio or video adapters, camera / camcorders, printer devices, keyboards, displays, etc. In several respects, the computing device provides interfaces, such as APIs or network services, which provide some or all of the data to other computing devices for further processing. Access to the interface can be opened and / or secured using any of a variety of technologies, such as by using a client authorization key, depending on the specific application requirements of this disclosure.

[0046] Network 140 may include a LAN (Local Area Network), WAN (Wide Area Network), telephone network (e.g., Public Switched Telephone Network (PSTN)), Session Initiation Protocol (SIP) network, wireless network, point-to-point network, star network, token ring network, hub network, wireless network (including protocols such as EDGE, 3G, 4G LTE, Wi-Fi, 5G, WiMAX, etc.), the Internet, etc. To ensure communication security, various authorization and authentication technologies can be used, such as username / password, Open Authorization (OAuth), Kerberos, SecurelD, digital certificates, etc. It should be understood that the network connections shown in the exemplary computing system 100 are illustrative, and any means of establishing one or more communication links between computing devices can be used. Authentication may be employed to ensure that client device 110, non-network client device 115, analytics server system 120, and animal monitoring device 130 are authorized to receive data and notifications traveling to and from each other. Users may use a username and password combination to log in and access data and / or notifications.

[0047] Figure 1B This is a bottom plan view of the animal monitoring device 130 that can be used in the animal health monitoring system and method disclosed herein, and Figure 1CThis is a side plan view of the animal monitoring device. The animal monitoring device in this example includes a platform 155 capable of carrying or receiving bedding material on top of the platform. In some examples, the platform has a bedding box 132, which is shown as being able to be placed on the upper surface of the platform. The bedding box is shown containing bedding material 134. The bedding box can be an off-the-shelf bedding box, can be custom-made for platform 155, or can be integrated with or coupled to the platform. The platform may be capable of carrying more than one type of bedding box. The platform is depicted as having a rectangular shape. However, the platform can be of any shape, such as square, rectangular, circular, triangular, etc.

[0048] Animal monitoring device 130 is depicted having four load sensors LC1, LC2, LC3, and LC4. It should be understood that the animal monitoring device can operate with three or more load sensors, and is not limited to four. Each of the four load sensors is associated with and separate from the platform 155, and receives pressure input independently of each other. In some examples, the platform may be triangular in shape and associated with three load sensors. The triangular shape allows the animal monitoring device to be easily placed in a corner of a room. The platform may be circular in shape and may have three or more load sensors. The platform is not limited to any type of shape and may have curves or straight lines, and rounded or sharp corners.

[0049] Animal monitoring device 130 may include processor 180 and memory 185. The processor and memory are capable of controlling a load sensor and receiving load data from the load sensor. The load data may be stored temporarily or permanently in the memory. Data communicator 190 is capable of transmitting the load data to another device. For example, the data communicator may be a wirelessly networked device employing an employee wireless protocol such as Bluetooth or Wi-Fi. The data communicator may send the load data to a physical remote device capable of processing the load data, such as… Figure 1A The analysis server system 120. The data communicator can also transmit data via a wired connection and may employ a data port such as a Universal Serial Bus (USB) port (including a USB-C port). Alternatively, the memory slot can accommodate a removable memory card on which payload data can be stored, and the removable memory card can then be physically removed and transferred to another device for uploading or analysis. In one embodiment, the processor 180 and memory 185 are capable of analyzing payload data without sending the payload data to a physically remote device such as the analysis server system.

[0050] Animal monitoring device 130 may include a power source 195. The power source may be a battery, such as a replaceable or rechargeable battery. The power source may be a wired power source plugged into a wall outlet. The power source may be a combination of a battery and a wired power source. Animal monitoring device 130 may be configured without a camera or image capture device and may not require the animal to wear an RFID collar.

[0051] Typically, a cat enters its litter box, finds a spot, defecates, covers its waste, and then leaves. While the cat is in the litter box, an animal health monitoring system can use one or more load sensors to track the cat's activity, measuring the cat's weight distribution and the total weight of the system. This data can be processed to identify specific cat characteristics, deriving features associated with the cat's behavior (e.g., defecation location, duration, movement patterns, entry force, exit force, event volatility, etc.). Various events can be identified based on these characteristics. In many implementations, various machine learning classifiers can be used to identify these events, as described in more detail herein. These events may include, but are not limited to, accidental triggering, human interaction, cat interaction outside the box, and cat interaction inside the box. Data analysis for classification can use normalization logic to normalize the raw data for classification. Aggregation logic can then be used with the normalized data to detect trends in data related to unique animals over time. These trends can be correlated with animal weight, average weight, weight trends, frequency patterns, time-of-day patterns, etc. The combined use of normalization techniques and aggregation logic can be referred to as set logic. The purpose of set logic is to transform the data output from the data model into a consistent set of measurements and provide smooth measurement trends across multiple time periods.

[0052] In one example, the load sensors may have a sampling rate associated with the data collected during interaction with the platform. The sampling rate may be set to a frequency or fall within a sampling rate range. Individual load sensors may be set to sample data at the same frequency as other sensors associated with the platform or at a different frequency. In one example, the sampling rate has a range of 2.5 Hz to 110 Hz. In another example, the sampling rate has a range of 20 Hz to 110 Hz. In yet another example, the sampling rate is 40 Hz. It should be understood that the load sensors may be separate and independent of each other, while being associated with the same platform. Data recorded or sampled by independent load sensors can allow for data analysis to determine unique characteristics of the animal's interaction with the device. For example, movement paths across the platform and into and out of the platform can be determined. Movement patterns and weight can be used to identify the animal. The platform may have a minimum of three load sensors or may have more than three load sensors, regardless of the platform's shape. In one example, the load sensors may include a full-bridge configuration with circular point contacts. In one example, the platform has X-axis dimensional measurements of 400mm to 600mm and Y-axis dimensional measurements of 250mm to 450mm. In another example, each load sensor may have its own maximum load capacity, which may be the same or different from each other. For example, each load sensor may have a maximum load capacity of 10kg. Overall, the individual load capacities of the load sensors limit the weight at which the device can be safely placed on the platform and function properly. The total weight placed on the platform may include the weight of the litter box, the litter, and the animal within the litter box. The maximum load capacity of the device may be less than the average weight of an adult. Therefore, in some examples, a person standing on the platform could cause it to malfunction. The maximum load capacity of the device may be selected based on the average weight of the animal (such as a cat) interacting with the device.

[0053] Figure 1D A bottom plan view of an animal monitoring device 131, which can be used in the animal health monitoring system and method disclosed herein, is depicted. The animal monitoring device may include a platform 156 that is substantially triangular in shape. The platform may have at least three or more load sensors. Animal monitoring device 131 may include some or all of the features and accessories described and depicted for animal monitoring device 130, such as a processor, memory, data communicator, and power supply. The triangular shape of the platform can be any type of triangle and is well-suited for installation in a corner of a room, such as a bathroom.

[0054] Figure 1EA bottom plan view of an animal monitoring device 133, which may be used in the animal health monitoring system and method of this disclosure, is depicted. The animal monitoring device may include a platform 158 that is substantially circular in shape. The platform may have at least three or more load sensors. The animal monitoring device 133 may include some or all of the features and accessories described and depicted for animal monitoring device 130, such as a processor, memory, data communicator, and power supply. The circular shape of the platform may be a circle of any size, or it may be an irregular circular shape including elliptical or oval shapes.

[0055] Platform 155 may have a foam pad adhered to the platform surface. The foam pad is used to connect the platform and container 132. The foam pad cushions the contact between the platform and the container and prevents the container from sliding on top of the platform. The foam pad may be embossed with designs such as logos or other features. The embossed design may have functional uses, such as anti-slip. The platform may also have physical buttons for user interaction with the platform, such as on / off, reset, pair, tare, and other buttons. After changes have been made (such as changing the bedding or replacing the bedding box itself), the weight of the bedding box and bedding can be reset using the tare button. The platform may include other hardware features that output information to the user, such as a display or speaker. Buttons can be used to navigate the interface on the display and select options. The platform may have adjustable feet on its bottom for adjusting the height of the platform and for stabilizing and leveling the platform on uneven surfaces.

[0056] Figure 2 A conceptual overview of events occurring within an animal health monitoring system according to exemplary aspects of this disclosure is shown. Event 200 may include error triggering, cat-in-the-box events, cat-out-of-the-box events, shovel events, and other events. Error triggering may indicate that some data was obtained from a load sensor, but no corresponding event occurred. Cat-in-the-box events may include excretion events (e.g., urination and / or defecation) and non-excretion events. When a cat-in-the-box event is detected, various characteristics of the cat can be determined. These characteristics include, but are not limited to, cat identification (cat ID), device balance, duration of the event, and cat weight. Cat-out-of-the-box events may include cat rubbing against the bedding box, cat standing on the edge of the bedding box, and / or cat standing on the top of the bedding box. Shovel events may include events in which a technician removes bedding and / or waste from the bedding box. Shovel events may include shoveling the bedding box, adding bedding to the bedding box, and moving the bedding box. For example, a user may pull the bedding box toward themselves and / or rotate the bedding box for easier access to all parts of the bedding box, thereby thoroughly removing waste. Other events may include the user moving the animal health monitoring system and / or bedding box. For example, the user may move the animal health monitoring system from one location to another, replace the bedding box located on top of the animal monitoring device, remove or replace the lid on the bedding box, and so on.

[0057] Activities related to bedding boxes can be represented as graphs with various peaks, valleys, flat points, and other characteristics, such as relative to... Figures 3A to 6B This is illustrated in more detail. For example, for a cat elimination event, there is typically an initial increase in weight when the cat enters the litter box; a period of movement as the cat moves within the litter box; a pause in activity during the elimination event; a second period of movement as the cat buries its excrement; and a decrease in the weight of the litter box when the cat leaves. As described in more detail herein, the flat points in the activity typically correspond to the actual elimination event. In some examples, the duration of a particular event provides an indication of the activity that occurred during the event. For example, most mammals take approximately 20 seconds to empty their bladder, and non-elimination events are generally shorter than urination events, which in turn are shorter than defecation events. Additionally, the change in the weight of the litter box after the event can be an indicator of the event that occurred, as urination events typically result in a greater weight increase than defecation events.

[0058] Activities may include a variety of events that can be identified and labeled using machine learning classifiers as described in more detail herein. Machine learning classifiers can be generally described as artificial intelligence (AI) models. Events may include, but are not limited to, the amount of movement a cat makes to find its litter box, the amount of time it takes to find its litter box, the amount of time it takes to prepare its litter box (e.g., digging in litter or other energy expended before defecating), the amount of time it takes to cover its litter box, the amount of effort (e.g., energy) it takes to cover its litter box, the duration of the litter box, the total duration of the event, the weight of the litter box, the animal’s movements during defecating (e.g., rubbing its rear end, shoving its rear end, etc.), step size / slope detection on a single load sensor during the litter box, the cat leaving its litter box, and movements and / or impacts involving the litter box.

[0059] Figures 3A to 3E The load signal for an event of a cat inside a box is shown according to an exemplary aspect of this disclosure. Figure 3A The image shows a signal 300 indicating a non-excretion event. Figure 3B The image shows a signal 310 indicating a urination event. Figure 3C The image shows a signal 320 indicating a defecation event. Figure 3D The image shows a signal 330 indicating a non-excretion event where a cat jumps into or out of the litter box. Figure 3E The image shows a signal 340 indicating an event in which the cat is partially inside the litter box during the covering action.

[0060] Figures 4A to 4C The load signal of a cat outside the box is shown according to an exemplary aspect of this disclosure. Figure 4A The image shows a signal 400 indicating an event where a cat rubs against the outside of the litter box. Figure 4B The image shows a signal 420 indicating an event that a cat is standing on the edge of the litter box. Figure 4C The image shows a signal 440 indicating an event that a cat is standing or sitting on top of the litter box.

[0061] Figures 5A to 5B A load signal for a removal event according to an exemplary aspect of this disclosure is shown. Figure 5A The image shows a signal 500 indicating a removal event. Figure 5B The image shows a signal 520 indicating a removal and relocation event.

[0062] Figures 6A to 6B A load signal for a movement event according to an exemplary aspect of this disclosure is shown. Figure 6A The image shows a signal 600 indicating the movement of the padding box. Figure 6B The image shows a signal 620 indicating a movement event of the measuring device.

[0063] Events can be conceptually divided into one or more phases for classification. For example, these phases may include a pre-excretion phase (e.g., entering, digging, searching), an excretion phase (e.g., urination, defecation), and a post-excretion phase (e.g., covering / leaving). Features can be built into the load data for each phase to identify specific behaviors occurring during that phase. Load data can be analyzed in both the time and signal domains. Time-domain features include, but are not limited to, mean, median, standard deviation, range, autocorrelation, etc. Time-domain features are created as input to machine learning classifiers. Frequency-domain features include, but are not limited to, median, energy, power spectral density, etc. Frequency-domain features are created as input to machine learning classifiers. For analysis related to a single excretion event, the Phase Separation Algorithm (PSA) can divide the event into three phases for classification. For analysis related to two excretions (e.g., both urination and defecation), PSA can divide the event into five phases, including a pre-excretion phase, a first excretion (urination) phase, a pre-excretion phase, a second excretion (defecation) phase, and a post-excretion phase.

[0064] Figure 7 The stages within an event according to an exemplary aspect of this disclosure are shown. For example... Figure 7As shown, event 700 may include three phases (e.g., phase 1, phase 2, and phase 3), measurements from each load sensor (e.g., load sensors 1 to 4), and the total load in the pad bin. In some embodiments, load data may be evaluated to determine the “flattest” point in the load data corresponding to the discharge event (e.g., phase 2), where data occurring before the flat point is phase 1 and data occurring after the flat point is phase 3. In several embodiments, a continuous sliding window may be used to analyze the load data. Sliding windows with the smallest variance differences (e.g., differences below a predetermined and / or dynamically determined threshold) are grouped together as potential flat points. The group with the largest sample size may be selected as the flat point of the event. In many embodiments, phases are determined based on the total load value, and the individual load sensor values ​​are divided into phases along the same time step defined by the total load. In some embodiments, events may be determined by analyzing the total load data and / or the load data of each individual load sensor. In many embodiments, events may be identified by identifying latent features in the load data of each load sensor and aggregating latent features to identify features within the total load data. The aggregation can be any mathematical operation, including but not limited to the sum and average of latent features.

[0065] In an implementation with two excretions and five stages, PSA can measure and detect two of the "flattest" points. For example, stages 2 and 4 would be the excretion stages. Stage 2 could be urination, and stage 4 could be defecation. Stage 1 could be the pre-excretion stage associated with the animal's movement before the first excretion. Stage 3 could be the second pre-excretion stage where the animal can move within the bedding box or otherwise adjust its position between stages 2 and 4. Stage 5 could be the post-excretion stage associated with the animal's removal from the bedding box. It should be understood that this technique can be used to identify and measure multiple excretion events, including more than two excretions. The total number of stages employed in PSA can be correlated with the number of excretions measured by adding a stage before and after the first and last excretions, and adding a stage between each excretion. Thus, PSA measuring a single excretion can employ three stages, and PSA measuring two excretions can employ five stages. In one implementation, a five-stage PSA can determine whether an animal interaction involves no excretion, only a single excretion, or two excretions.

[0066] In many implementations, one or more machine learning classifiers may be used to analyze payload data to identify and / or label events within the payload data. Based on these labels, events and / or animals may be classified. It will be apparent to those skilled in the art that a variety of machine learning classifiers may be utilized, including (but not limited to) decision trees (e.g., random forests), k-nearest neighbors, support vector machines (SVMs), neural networks (NNs), recurrent neural networks (RNNs), convolutional neural networks (CNNs), and / or probabilistic neural networks (PNNs). RNNs may also include (but are not limited to) fully recurrent networks, Hopfield networks, Boltzmann machines, self-organizing maps, learned vector quantization, simple recurrent networks, echo-state networks, long short-term memory networks, bidirectional RNNs, hierarchical RNNs, stochastic neural networks, and / or genetic-scale RNNs. In several implementations, combinations of machine learning classifiers may be utilized. More specific machine learning classifiers (when available) and general machine learning classifiers at other times may further increase the accuracy of predictions.

[0067] Figure 8 A flowchart of a method 800 (or process) for classifying animal behavior according to an exemplary aspect of this disclosure is shown. Although the method is described with reference to the flowchart, it should be understood that many other methods may be used to perform the actions associated with the method. For example, the order of some boxes may be changed, some boxes may be combined with other boxes, one or more boxes may be repeated, and / or some of the boxes described are optional. The method may be executed by processing logic, which may include hardware (circuit, special-purpose logic, etc.), software, or a combination of both. The method or process may be implemented as instructions on a machine, wherein the instructions are included on at least one computer-readable medium or a non-transitory machine-readable storage medium.

[0068] according to Figure 8Load data 810 can be obtained, for example, from one or more load sensors in an animal health monitoring system as described herein. In further details, stage data 812 can be identified, including stage data such as the search stage, excretion stage, and / or covering stage as described herein. However, it should be noted that this stage data is provided by way of example only, as different stages may be identified for different animals as appropriate. In some examples, time-domain features 814 and / or frequency-domain features 816 can be identified. For example, load data may include information in the time domain, frequency domain, or both. In some embodiments, load data can be transformed from time-domain data to frequency-domain data. For example, various techniques (such as Fourier transform) can be used to transform time-domain data to frequency-domain data. Similarly, various techniques (such as inverse Fourier transform) can be used to transform frequency-domain data to time-domain data. In some embodiments, time-domain features and / or frequency-domain features can be identified based on specific peaks, troughs, and / or flat points within the time-domain data and / or frequency-domain data as described herein.

[0069] In relative to Figure 8 In further details, feature 818 may be selected, for example, from stage data, time-domain features, and / or frequency-domain features of individual and / or all load sensors. In some implementations, feature 820 may be classified, for example, using a machine learning classifier, and in some examples, features may be classified simultaneously by a machine learning classifier. Classifying events may include determining labels for identifying features and confidence measures indicating the probability that the labels correspond to the basic facts of the event (e.g., the probability that the label is correct). These labels may be determined based on features, stages, and / or a variety of other data.

[0070] The established features can be used to classify behaviors using one or more machine learning classifiers as described herein. For example, a variety of features can be established or created in the time and / or frequency domains. These features include, but are not limited to, standard deviation of load, length of flat points, cross counts of means, counts of unique peaks, counts of different load values, ratios of different load values ​​to event durations, counts of maximum load variations across individual sensors, percentage of medium-load bins, percentage of high-load bins, high-load bin volatility, high-load bin variance, autocorrelation function hysteresis or delay, curvature, linearity, peak counts, energy, minimum power, power standard deviation, maximum power, maximum variance offset, maximum KL divergence, KL divergence time, spectral density entropy, autocorrelation function differentiation, and / or changes in the autoregressive model. Therefore, behaviors can be classified based on their relevance to the features being classified. For example, the selected features can be used as input to a machine learning classifier to classify behaviors. The classified behaviors can include labels indicating the type of behavior and / or confidence measures indicating the likelihood that the labels are correct. The machine learning classifier can be trained on a variety of training data indicating animal behavior and basic fact labels with features as input.

[0071] In such Figure 8 In further details, event 822 can be categorized, such as based on the created characteristics and / or phase data. In some implementations, events can be categorized based on a confidence metric indicating the likelihood that one or more events have been correctly classified. For example, events can be classified as excretion events, shovel events, cat sitting on the litter box events, and / or any of a variety of other events as described herein. In further details, events can cause changes in the overall state of the animal health monitoring system. For example, adding litter, changing litter, and shoveling events can cause changes in the total weight of the litter box. In these cases, the animal health monitoring system can be recalibrated to maintain the accurate performance of the animal health monitoring system.

[0072] A transmittable notification 824 may include notifications relating to animal behavior generated based on categorization events and / or historical events of the animal. In some embodiments, the notification may be generated based on events of other animals in the same cohort as the animal. The notification may indicate that an event has occurred and / or may indicate one or more inferences about the animal. For example, an animal's urination behavior may be tracked over time, and if urination activity increases or decreases (the decrease may be due to straining during urination or increased non-excretory access to the bedding), a notification may be generated indicating that the animal may have a urinary tract infection or other illness requiring veterinary attention. However, any behavior and / or characteristic of the animal (such as weight) may be used to trigger notification generation. In some embodiments, the notification is transmitted once a threshold amount of data and / or events has been determined. The notification may be transmitted to a client device associated with the animal's owner and / or veterinarian, as described herein. In several embodiments, the notification provides an indication requesting user confirmation that the detected event is correct. In this way, the notification can be used to obtain a basic truth label for the event, which can be used to train and / or retrain one or more machine learning classifiers.

[0073] As previously described, load data can be analyzed at the stage level, by total load, individual load of each load sensor, and / or via a stage separation algorithm that segments the load data into stages. Exemplary stages may include pre-excretion (e.g., entering, searching, digging), excretion (e.g., urinating, defecating), and post-excretion (e.g., covering, leaving). In addition to these characteristics, animal behavior and location can also be determined. In several embodiments, the animal's location within the bedding box can be determined based on the animal's center of gravity position within the bedding box at different times during the event. By tracking the animal's center of gravity, the animal's location within the bedding box can be determined for each stage and / or each characteristic within the event.

[0074] Figure 9AAn example of position tracking 900 for an animal's movement path according to an example of this disclosure is shown. The animal's movement path within a litter box can be described from the inlet to the outlet of the litter box. The animal's center of gravity can be used to track the movement path. In this example, an animal health monitoring system can be used, which includes an animal monitoring device 130, which includes a platform 155 and a plurality of load sensors LC1, LC2, LC3, and LC4, each load sensor being located near a corner of the litter box on the platform. The animal monitoring device carries a litter box (not shown) containing litter on it. For convenience, a coordinate system can be defined, where the center of the platform (which may be aligned with the center of the litter box) is defined as (0,0), the first corner where LC1 is approximately located is defined as (-1,1), the second corner where LC2 is approximately located is defined as (-1,-1), the third corner where LC3 is approximately located is defined as (1,1), and the fourth corner where LC4 is approximately located is defined as (1,-1).

[0075] In this example, the initial center of gravity of the animal health monitoring system can be calculated based on the tare weight (empty weight) of the animal health monitoring device, which is loaded with bedding. As the animal enters the bedding box, each load sensor obtains a different load measurement depending on the animal's position within the box. At a given time, the animal's center of gravity can be calculated based on the measurements from each load sensor. Figure 920 illustrates various positions of the animal's center of gravity when in the bedding box placed on top of the animal monitoring device, including approximate entry and exit points. Because each animal has its own unique personality, habits, and daily activities, the general movement of an animal during a specific category of events is often unique to that animal. Thus, animal movement data can be used to identify the animal's signature during a specific event.

[0076] In addition to the animal's movement patterns during a specific event, a variety of other characteristics of the event can be used to classify the event and / or identify a specific animal. These characteristics include, but are not limited to, the animal's weight, the time when the animal typically engages in a particular type of event, the animal's location during one or more phases of the event, covering behavior (e.g., covering in place, leaving and returning to the litter box to cover, covering while standing halfway in the litter box, scratching the litter box, etc.), climbing over the edge of the litter box or jumping into the litter box, the total duration of activity within the box, preference for placing the litter box in one unit rather than another in a multi-unit environment, the typical weight of the excrement, the time of entry / exit before excretion, the time spent digging before / after excretion, the force used to cover the excrement, the speed of paw movement during covering, movement patterns within the litter box (e.g., clockwise and / or counterclockwise movement), consistency of excrement spot selection, and the order in which cats enter the box in a multi-cat household.

[0077] Many pet owners own multiple animals that share the same litter box. Therefore, the animal health monitoring system disclosed herein can be adjusted or modified to differentiate between multiple animals using the same litter box. Based on this, in Figure 9B and Figure 9C An example is provided demonstrating how animals can be identified based on their behavior even when multiple animals are using the same bedding box. For instance, a machine learning classifier can select multiple features associated with a cat's behavior within the box. Furthermore, principal component analysis (PC1, PC2, etc.) as a dimensionality reduction technique can be performed on all features to create the first two principal components composed of these features. (See the examples provided.) Figures 9B to 9C The diagrams at positions 940 and 960 illustrate PC1 and PC2, separated by individual cats, demonstrating how features can be used to cluster cats and assign animal identifiers. Data processing for analyzing animal identification data from load sensors can employ normalization logic. Normalization logic can be used to resolve data conflicts between different types of events. Normalization logic can take input from the user to correct the output of the data analysis. For example, the user can correct the cat's identity.

[0078] Figure 10 A flowchart of a method 1000 (or process) for animal identification according to an example of this disclosure is shown. Although referenced... Figure 10 The flowchart shown illustrates the method; however, it should be understood that many other methods may be used to perform the actions associated with this method. For example, the order of some boxes may be changed, some boxes may be combined with other boxes, one or more boxes may be repeated, and / or some of the boxes described are optional. The method may be executed by processing logic, which may include hardware (circuit, special-purpose logic, etc.), software, or a combination of both. The method may be implemented as a method and executed as instructions on a machine, wherein the instructions are included on at least one computer-readable medium or a non-transitory machine-readable storage medium.

[0079] According to method 1000, load data 1010 can be obtained, stage data 1012 can be determined at box 1012, and event 1014 can be determined, as described herein. Animal paw prints 1016 (or signals) can be determined and used to identify typical movement patterns of animals during specific types of events. Animal movement patterns can be determined based on a variety of characteristics of the animal's center of gravity movement during the event, including but not limited to coverage distance, speed, acceleration, direction of movement, alignment, distance from the bedding box entry point to the center of the bedding box, excretion point, resting point, and preferred quadrant of the bedding box. In some embodiments, an animal's preference for a particular quadrant can be determined based on the percentage of total observations in each quadrant and the percentage of load observations in each quadrant relative to the total number of load samples measured. In other embodiments, animal traits can be determined by identifying and / or computing one or more features within the movement data as input to one or more machine learning classifiers.

[0080] In further detail, animals can be identified, for example, based on animal traits, identified events, and / or one or more characteristics of the events 1018. Animal behavior models 1020 can be generated, which in some examples may indicate animal traits in relation to a variety of events. For example, an animal behavior model may indicate the event, the frequency of the event, the animal's traits in relation to the event, the animal's preferred behavior during the event, the characteristics of the event and / or the animal, and / or any other relevant or available information, such as those also described herein.

[0081] Method 1000 may also include transmitting a notification 1022. This notification may be generated and / or transmitted based on a specific animal performance event. The notification may be sent to a client device and may include the animal's instructions and / or any other information as described herein. Multiple notifications and technologies for delivering notifications are possible. For example, notifications indicative of multiple insights into the behavior of their pets may be sent to users. These notifications may be sent on any schedule (e.g., daily, weekly, monthly, etc.) and / or based on the fulfillment of specific notification thresholds. Notifications may include a summary of any animal monitoring devices in the same household, the animal's preference for different defecation locations for urination or defecation, time-of-day reports indicating the animal's typical daily activities, relevant instructions on optimal bedding box maintenance times based on animal activity, and / or any other appropriate insights.

[0082] Notification thresholds can be based on any aspect of the animal that may require additional analysis, such as weight loss or gain exceeding a threshold amount within a specific timeframe, an increase or decrease in excretion events, more or less frequent access to the excretion area, changes in excretion routines, and / or any other factor or combination of factors indicating potential health problems, as described herein. A variety of animal characteristics can be provided, as described in more detail below. These characteristics may include, but are not limited to, age, sex, reproductive status, and / or physical condition. These factors can be used to establish notification thresholds and / or to provide insights when an animal reaches specific thresholds for weight change, access, and / or excretion frequency. For example, the threshold for a kitten in ideal physical condition would differ from the threshold for an underweight older cat.

[0083] These notifications can provide indications of potential problems with a cat's health and / or emotional state. For example, fluctuations in weight and visitation frequency can be early indicators of many disease states, such as feline lower urinary tract infections, bladder stones, bladder crystals, kidney disease, diabetes, hyperthyroidism, feline spontaneous cystitis, digestive problems (IBD / IBS), and arthritis and / or emotional health issues such as stress, anxiety, and cognitive decline / dysfunction. For many animals, changes in health or behavioral status can go unnoticed until symptoms become extremely severe. Notifications provided by animal health monitoring systems can provide early indicators of changes in an animal's health or behavior. Animal health monitoring systems, as described herein, can help identify these potential problems at an early stage. For example, some problems or conditions can be defined by stages, such as stages 1 through IV. In this example, notifications can be sent to the pet owner during an earlier stage (e.g., stage I or stage II) to allow treatment to be implemented before the animal's overall health is more severely affected (e.g., in stage III or stage IV).

[0084] As mentioned, in some examples, animal health monitoring systems can be used in environments with multiple animals. These animals may have different weights and / or they may have similar weights (e.g., their weights may overlap). Existing systems that use animal weight to identify animals generally perform poorly in these systems because weight is not a unique indicator for a particular animal. Instead, animal health monitoring systems as described herein can use a variety of models (such as feature-based models, activity models, and combinations of models) to uniquely identify the animals using the animal health monitoring system.

[0085] Figure 11The performance of various classifiers or classification models according to exemplary aspects of this disclosure is illustrated. As shown in the table at 1100, a hybrid model that simultaneously analyzes the characteristics of an event and the location of an animal during the event can be comparable to or even superior to a single model for all numbers of cats and all categories of overlap weight. However, it should be noted that, depending on the specific application of the embodiments provided by this disclosure, one or more models may be used to identify animals and events.

[0086] It should be understood that all the disclosed methods and processes described herein can be implemented using one or more computer programs, components, and / or program modules. These components may be provided as a series of computer instructions on any conventional computer-readable or machine-readable medium, including volatile or non-volatile memories such as RAM, ROM, flash memory, magnetic or optical disks, optical storage, or other storage media. The instructions may be provided as software or firmware and / or may be implemented, wholly or partially, in hardware components such as ASICs, FPGAs, DSPs, or any other similar devices. The instructions may be configured to be executed by one or more processors, which, in executing the series of computer instructions, perform or facilitate the execution of all or part of the disclosed methods and processes. As those skilled in the art will appreciate, the functionality of the program modules may be combined or distributed as needed across various aspects of this disclosure.

[0087] Figures 12 to 14 Flowcharts of methods 1200, 1300, and 1400 (or processes) for monitoring animal health according to exemplary aspects of this disclosure are shown. Although referenced... Figures 12 to 14 The flowcharts shown illustrate these methods; however, it should be understood that many other methods may be used to perform the actions associated with this method. For example, the order of some boxes may be changed, some boxes may be combined with other boxes, one or more boxes may be repeated, and some of the boxes described are optional. The method may be executed by processing logic, which may include hardware (circuit, special-purpose logic, etc.), software, or a combination of both. The method may be implemented as a method and executed as instructions on a machine, wherein the instructions are included on at least one computer-readable medium or a non-transitory machine-readable storage medium.

[0088] Therefore, according to Figure 12A method 1200 for monitoring animal health under the control of at least one processor may include obtaining 1210 load data from a plurality of load sensors associated with a platform on which bedding material is carried. Each of the plurality of load sensors may be separate from each other and receives pressure input independently. In further detail, the method may include determining 1212 whether the load data originates from an interaction between the animal and the bedding material; if the load data determines that the interaction with the bedding material is due to animal interaction, identifying 1214 an animal-related behavioral attribute; classifying 1216 the animal behavioral attribute as an animal classification event using a machine learning classifier; and identifying 1218 a change in the animal classification event compared to previously recorded animal-related events.

[0089] In some examples, the categorization of animal behavior may include one or more of in-box events, urination events, defecation events, or non-excretion events. Method 1200 may also include associating variations in animal categorization events with physical, behavioral, or mental health issues related to the animal. In other examples, physical health issues may be animal diseases. In other examples, animal diseases may be selected from feline diseases such as urinary tract diseases, kidney disease, diabetes, hyperthyroidism, spontaneous cystitis, digestive problems, and arthritis. In some examples, mental health issues may be selected from anxiety, stress, and cognitive decline. In other examples, behavioral problems may be out-of-box excretion. In other examples, determining whether the load data originates from animal interaction further determines whether the load data originates from animal interaction, human interaction, error triggering, or accidental interaction.

[0090] Method 1200 may also include identifying animals based on load data. In some examples, identifying an animal distinguishes it from at least one other animal interacting with the platform. The method may also include generating notifications indicating changes in animal classification events. In other examples, notifications are generated after parameters associated with device events meet thresholds. In other examples, the method may not include or communicate with any camera or image capture device and may not perform visual image recognition. In some examples, classifying animal behavioral attributes also includes analyzing load data from multiple load sensors to measure one or more of the following: (i) the weight of the bedding bin positioned on the platform, (ii) the weight distribution of the animal, (iii) the location of the event, (iv) the duration of the event, (v) movement patterns, (vi) entry forces, (vii) exit forces, or (viii) the volatility of animal interactions. In other examples, classifying animal behavioral attributes also includes analyzing load data from multiple load sensors to identify or measure one or more of the following: (i) an animal entering a bedding box on a platform, (ii) the amount of movement an animal makes when selecting a particular excretion location, (iii) the amount of time taken to select a particular excretion location, (iv) the amount of time taken to prepare (e.g., dig) the excretion location before excretion, (v) the amount of time taken to cover the excrement, (vi) the energy taken to cover the excrement, (vii) the duration of excretion, (viii) the total duration of the device event from the animal's entry to its exit, (ix) the weight of the excrement, (x) the animal's movement during excretion, (xi) step length / slope detection on a single load sensor during excretion, (xii) the animal leaving the positioned bedding box, or (xiii) one or more movements or impacts involving the bedding box.

[0091] In some examples, classifying animal behavioral attributes also includes analyzing load data from multiple load sensors in both the time and frequency domains. In other examples, one or more time-domain features include the mean, median, standard deviation, range, or autocorrelation created as input to a machine learning classifier. In other examples, one or more frequency-domain features include the median, energy, or power spectral density created as input to a machine learning classifier. In some examples, the selected features are chosen from both the time and frequency domains, and the selected features are one or more of the following: (i) standard deviation of the load, (ii) length of the flat point, (iii) cross count of the mean, (iv) count of unique peaks, (v) count of different load values, (vi) ratio of different load values ​​to event duration, (vii) count of maximum load variation in each sensor, (viii) percentage of medium load bins, (ix) percentage of high load bins, (x) high load bin volatility, (xi) high load bin variance, and (xii) autocorrelation function. Lag or delay, (xiii) curvature, (xiv) linearity, (xv) peak count, (xvi) energy, (xvii) minimum power, (xviii) power standard deviation, (xix) maximum power, (xx) maximum variance shift, (xxi) maximum KL divergence, (xxii) KL divergence time, (xxiii) spectral density entropy, (xxiv) autocorrelation function derivative or (xxv) autoregressive model variation; and wherein animal interactions are classified and / or animal identification is determined based on using the selected features as input to a machine learning classifier.

[0092] In some examples, classifying animal behavioral attributes in this method and other methods 1200 further includes analyzing load data from multiple load sensors at the stage level by (i) total load, (ii) individual load for each load sensor, and (iii) a stage separation algorithm that segments the load data into stages. In other examples, the stage separation algorithm that segments the load data into stages includes at least three stages, including pre-excretion, excretion, and post-excretion. In other examples, the method also includes determining the animal's position within a litter box positioned on a platform. In some examples, the animal's position within the litter box is based on the animal's center of gravity position within the litter box at different times during animal interaction. In other examples, the method also includes tracking the animal's center of gravity to determine the animal's position within the litter box for each stage and / or each feature within the animal interaction.

[0093] In some examples, classifying animal behavioral attributes also includes analyzing load data from multiple load sensors to determine the animal's movement pattern, which includes one or more of the following: (i) coverage distance, (ii) speed, (iii) acceleration, (iv) direction of movement, (v) alignment, (vi) distance from the entry point into the bedding bin positioned on the platform to the center of the bedding bin, (vii) excretion location, (viii) resting location, or (ix) the preferred quadrant of the bedding bin. In other examples, the preferred quadrant is determined based on the percentage of total observations in each quadrant and the percentage of load observations in each quadrant relative to the total load sample. In other examples, method 1200 also includes generating an animal behavior model for a specific animal, including identifying one or more of the following: (i) the device events of the specific animal, (ii) the frequency of the device events, (iii) the traits of the specific animal during the device events, (iv) the preferred behavior of the specific animal during the device events, or (v) the device events and / or the characteristics of the specific animal.

[0094] Multiple user interfaces are available to ensure the proper installation, configuration, and use of the animal health monitoring system. These user interfaces can provide instructions to the user, request information from the user, and / or provide insights into the behavior and potential problems of one or more animals.

[0095] When establishing an animal health monitoring system, the initialization and positioning of the animal monitoring device are crucial to ensuring the accuracy of the collected load data. In some embodiments, the animal monitoring device operates best in an indoor climate-controlled environment without direct sunlight. In several embodiments, the animal monitoring device should be placed at least one inch away from all walls or other obstructions, as insufficient space may cause the device to become stuck, interfering with data or readings. Additionally, the animal monitoring device should be located at a sufficient distance from high-vibration objects (such as washing machines and dryers) or high-traffic areas, as vibration can cause erroneous and / or inaccurate readings from the weight sensor. In many embodiments, the animal monitoring device operates best on smooth, level, hard surfaces, as soft or uneven surfaces can affect the accuracy of the load sensor. In one embodiment, the animal monitoring device has adjustable feet to level it on uneven surfaces. In many embodiments, the animal monitoring device can be slowly introduced to the animal to improve its integration into the environment. For example, the animal monitoring device can be placed in the same room as the bedding box for several days to allow the animal to acclimate to its presence. Once the animal is accustomed to the presence of the animal monitoring device, the intensity of the device can be reduced to allow the animal to adjust to the subtle sounds and lights it may emit. After the animal is comfortable with the device, a litter box can be placed on top of it. Adding new litter to the litter box encourages the animal to use it.

[0096] In some implementations, multiple user interfaces are used for configuring the animal health monitoring system. These user interfaces may include a user interface for initiating the animal monitoring device setup process, a user interface for initiating the network setup process, a user interface for connecting to the animal monitoring device via Bluetooth during the setup process, a user interface for confirming the Bluetooth connection to the animal monitoring device during the setup process, a user interface for connecting the animal monitoring device to a local area network, a user interface indicating that the animal monitoring device is ready for use, a user interface for physically locating the animal monitoring device and bedding box, and / or a user interface for confirming the completion of the setup process.

[0097] A profile can be generated for each animal. This profile can be used to establish baseline characteristics for each animal and track the animal's behavior and characteristics over time. This may include tracked weight, number and type of events, waste type, time of day for each event, and / or any other data as described herein.

[0098] In some implementations, a user interface is used for creating animal profiles. Examples of user interfaces for creating animal profiles include a user interface for a start screen of the animal profile creation process, a user interface for an introductory screen of the animal profile creation process, a user interface for entering the animal's name, a user interface for entering the animal's sex, a user interface for entering the animal's reproductive status, a user interface for an introductory screen explaining the current physical condition of the captured animal, a user interface for checking the animal's ribs, a user interface for checking the animal's profile, a user interface for checking the animal's waist, a user interface for an end screen of the animal profile creation process, a user interface for the type or brand of bedding used (including the characteristics of the bedding), a user interface for the type of bedding used, and / or a user interface for feeding the animal's diet.

[0099] Each cat is unique and exhibits unique behaviors. Animal health monitoring systems can utilize multiple machine learning classifiers to track and differentiate multiple animals without the need for additional collars or gadgets. In some implementations, information about specific events, such as identification of which cat has used the litter box, can be requested from the user. This information can be used to confirm the identity of the animal associated with the specific event, which can be used to retrain the machine learning classifier and improve the accuracy of future results. For example, if an animal's behavior and weight change, the system can request confirmation of which animal is associated with a particular event so that the system continues to provide the best available insights. In other implementations, fewer event confirmations can be provided when animals in a multi-animal environment have different weights. In many implementations, if the animals are of roughly the same weight, placing each cat and animal monitoring device in a separate room reduces the number of confirmation requests. In several implementations, once the system has established a unique profile for a particular animal (e.g., after a threshold number of confirmations), the frequency of future confirmation requests can be reduced.

[0100] In some implementations, a user interface for tagging events may be used. The user interface may include a user interface for displaying notifications, a user interface for requesting additional information about the event, a user interface for requesting identification of animals involved in the event, and a user interface for displaying requested information associated with the event.

[0101] As described in this article, animal characteristics and behaviors can be tracked and analyzed over time. Data can be analyzed within any time frame (such as, but not limited to, 24 hours, 48 ​​hours, one week, two weeks, one month, etc.). Analysis of animal behavior and characteristics over time can be used to identify when changes in the animal's typical condition occur, which may be indicators of adverse events requiring additional diagnosis or treatment.

[0102] In some implementations, a user interface for tracking animal behavior may be used. Examples of user interfaces for tracking animal behavior include those displaying animal weight over a one-week period, over a thirty-day period, over a one-year period, displaying the number of times animal weight was measured on a specific day, over a thirty-day period, over a one-year period, displaying the number of times animal weight was measured, displaying the number of times animal weight was measured over a one-year period, displaying the number of events at three different bedding boxes over a one-week period, displaying the number of events at one bedding box over a one-week period, displaying an indication of the type of event occurring at one bedding box, displaying the number of events at one bedding box over a one-week period, and / or displaying the number of excretion events at multiple bedding boxes. In one example, a household or other location may include multiple devices with bedding boxes having implementations of this technology. The household may also include more than one animal using the device. Data from multiple devices may be aggregated to provide insights into the behavior of each animal at the household level.

[0103] As described in this article, various notifications indicating potential health problems in animals can be provided based on changes in animal behavior. However, if animal monitoring devices have become misaligned or improperly calibrated, these indicated changes may be false alarms. In such cases, the correct operation of the monitoring event should be confirmed before determining whether additional attention should be given to the animal to ascertain whether any adverse health changes have occurred.

[0104] In some implementations, a user interface is used for expert advice notification. The user interface may include a user interface displaying a notification indicating that the cat should be monitored due to weight loss, a user interface requesting confirmation that the animal monitoring device is correctly configured, a user interface requesting additional information about the cat's eating behavior, a user interface requesting additional information about the cat's appearance, a user interface requesting additional information about the cat's excretion, and / or a user interface providing guidance to contact a veterinarian when changes in the cat's behavior or condition are of concern.

[0105] In addition to animal behaviors and activities as described in this article, animal health monitoring systems also track and record a variety of non-animal activities. Multiple user interfaces are available to provide insights into these animal and non-animal behaviors. For example, insights into typical animal behaviors can suggest the ideal time for cleaning and / or maintaining bedding boxes.

[0106] In some implementations, a user interface for animal behavior analysis is used. Examples of user interfaces for animal behavior analysis include a user interface that displays the general behavior of two animals over a period of time, a user interface that displays the litter box preferences of two animals over a period of time, a user interface that displays the time-of-day behavior patterns of two animals over a period of time, a user interface that compares the time-of-day behavior patterns of two animals and a user over a period of time, and / or a user interface that compares the time-of-day excretion behavior of two animals and user maintenance events over a period of time.

[0107] according to Figure 13 The method 1300 for monitoring animal health may include obtaining 1310 load data from an animal monitoring device, the animal monitoring device including three or more load sensors associated with a platform on which contained bedding material is carried, wherein each of the three or more load sensors is separate from each other and independently receives pressure input from the platform, wherein the three or more load sensors individually sample the load at a frequency of 2.5 Hz to 110 Hz. In further details, the method may include identifying 1312 animal-related behavioral attributes if it is determined from the load data that the interaction with the contained bedding material is due to the interaction between the animal and the contained bedding material. The method may include classifying the animal behavioral attributes 1314 as animal classification events using a machine learning classifier.

[0108] In some examples, the method 1300 claimed in the claims includes the animal monitoring device transmitting load data, animal sorting events, or both from a data communicator to a client device via a network. In some examples, method 1300 also includes identifying changes in animal sorting events compared to previously recorded animal sorting events or patterns of previously recorded animal sorting events. In one example, three or more load sensors individually sample the load at 20 Hz to 110 Hz. In one example, three or more load sensors individually sample the load at 30 Hz to 80 Hz. In one example, three or more load sensors individually sample a load of up to 10 kg. In one example, the animal monitoring device includes four load sensors. In one example, the platform has a rectangular shape, and the four load sensors are each positioned at a different corner of the platform. In one example, the animal monitoring device has three load sensors, and the platform is triangular in shape. In one example, the platform has a triangular shape, and the three load sensors are each positioned at a different corner of the platform. Figure 14A method 1400 for monitoring animal health under the control of at least one processor may include obtaining 1410 load data from an animal monitoring device including three or more load sensors associated with a platform on which contained bedding material is carried, wherein each of the three or more load sensors is separate from each other and receives pressure input from the platform independently. In further details, the method may include identifying 1412 animal-related behavioral attributes if it is determined from the load data that the interaction with the contained bedding material is due to the interaction between the animal and the contained bedding material. The method may also include classifying the animal behavioral attributes 1414 as animal classification events using a machine learning classifier, including analyzing the load data via a stage separation algorithm, wherein animal classification events include animal excretion, and wherein the stage separation algorithm is capable of classifying multiple discrete animal excretions occurring at multiple discrete animal excretions.

[0109] In some examples, method 1400 also includes multiple discrete animal excretions, which include both urination and defecation events. Multiple discrete animal excretions may include a first excretion and a second excretion, and the animal classification event further includes a pre-first excretion phase preceding the first excretion and a pre-second excretion phase preceding the second excretion. The animal classification event may also include a post-excretion phase occurring after the second excretion. Method 1400 may also include classification using a phase separation algorithm to determine that multiple discrete animal excretions did not exist during the animal interaction, either in the form of no animal excretion or only a single animal excretion. Method 1400 may also include identifying changes in the animal classification event compared to previously recorded animal classification events. Method 1400 may also include associating changes in the animal classification event with animal-related physical, behavioral, or mental health problems.

[0110] In one example of method 1400, physical health problems are selected from feline diseases such as urinary tract diseases, kidney disease, diabetes, hyperthyroidism, spontaneous cystitis, digestive problems, or arthritis, and mental health problems are selected from anxiety, stress, or cognitive decline. In one example, identifying changes in animal categorization events includes identifying changes in animal excretion. In one example, method 1400 also includes generating a notification indicating a change in animal categorization events, wherein the notification is generated after a parameter associated with the device event meets a threshold. In one embodiment, the method also includes identifying the animal based on load data and distinguishing the animal from at least one other animal interacting with the contained bedding.

[0111] In one example of method 1400, classifying animal behavioral attributes also includes analyzing load data from three or more load sensors based on the animal's interaction with the contained bedding to measure the animal's weight distribution, location of events, duration of events, movement patterns, entry forces, exit forces, variability in animal interactions, or combinations thereof. In another example of method 1400, animal behavioral attributes also include analyzing load data from three or more load sensors to identify or measure an animal entering a bedding bin containing the contained bedding, the amount of movement an animal makes when selecting a particular defecation location, the amount of time taken to select a particular defecation location, the amount of time taken to prepare (e.g., dig) a defecation location, the amount of time taken to cover the excrement, the energy consumed in covering the excrement, the duration of defecation, the total duration of the device event from the animal's entry to its exit, the weight of the excrement, the animal's movement during defecation, step length / slope detection on a single load sensor during defecation, the animal leaving the located bedding bin, one or more movements or impacts involving the bedding bin, or combinations thereof. In one example of method 1400, classifying animal behavioral attributes further includes analyzing load data from three or more load sensors in both the time domain (based on time-domain features) and the frequency domain (based on frequency-domain features) based on the animal's interaction with the contained bedding. The time-domain features include mean, median, standard deviation, range, autocorrelation, or combinations thereof, and are created as one or more inputs to a machine learning classifier. The frequency-domain features include median, energy, power spectral density, or combinations thereof, and are also created as one or more inputs to a machine learning classifier. In another example of method 1400, classifying animal behavioral attributes further includes analyzing load data from three or more load sensors based on the animal's interaction with the contained bedding to determine the animal's movement patterns, including coverage distance, velocity, acceleration, direction of movement, alignment, distance from the entry point into the bedding bin positioned on the platform to the center of the bedding bin, excretion location, resting location, preferred quadrant of the bedding bin, or combinations thereof.

[0112] In one example, method 1400 also includes generating animal behavior models for each animal based on animal classification events unique to each animal. In one example, three or more load sensors individually have sampling rates from 2.5 Hz to 110 Hz. In one example, three or more load sensors individually have a maximum load capacity of up to 10 kg.

[0113] It should be understood that all the disclosed methods and processes described herein can be implemented using one or more computer programs, components, and / or program modules. These components may be provided as a series of computer instructions on any conventional computer-readable or machine-readable medium, including volatile or non-volatile memories such as RAM, ROM, flash memory, magnetic or optical disks, optical storage, or other storage media. The instructions may be provided as software or firmware and / or may be implemented, wholly or partially, in hardware components such as ASICs, FPGAs, DSPs, or any other similar devices. The instructions may be configured to be executed by one or more processors, which, in executing the series of computer instructions, perform or facilitate the execution of all or part of the disclosed methods and processes. As those skilled in the art will appreciate, the functionality of the program modules may be combined or distributed as needed across various aspects of this disclosure.

[0114] Although this disclosure has been described in certain specific aspects, many additional modifications and variations will be apparent to those skilled in the art. Specifically, any of the various processes described above may be performed in an alternative sequence and / or in parallel (on the same or different computing devices) to achieve similar results in a manner more suited to the requirements of a particular application. Therefore, it should be understood that this disclosure may be practiced in ways different from those specifically described without departing from the scope and spirit of this disclosure. Consequently, aspects of this disclosure should be considered illustrative rather than restrictive in all respects. It will be apparent to those skilled in the art that several or all aspects discussed herein that are considered suitable for a particular application of this disclosure may be freely combined. Throughout this disclosure, terms such as “advantageous,” “exemplary,” or “preferred” indicate elements or dimensions that are particularly suitable (but not essential) for this disclosure or embodiments thereof, and modifications to these elements or dimensions may be made wherever a skilled observer deems it appropriate, except where expressly required. Therefore, the scope of this disclosure should not be determined by the illustrated embodiments but rather by the appended claims and their equivalents.

Claims

1. A method for monitoring the health status of an animal, comprising: Load data is obtained from three or more load sensors of an animal monitoring device, the three or more load sensors being associated with a platform on which the contained bedding material is carried, wherein each of the three or more load sensors is separate from each other and receives pressure input from the platform independently, wherein the three or more load sensors individually sample the load at a frequency of 2.5 Hz to 110 Hz. If it is determined from the load data that the interaction with the contained bedding is due to the interaction between the animal and the contained bedding, then animal behavioral attributes related to the animal are identified; The animal behavioral attributes are classified into animal classification events using a machine learning classifier; as well as The animal's movement patterns are tracked based on the load data.

2. The method of claim 1, wherein the animal monitoring device transmits the load data, the animal classification event, or both from the data communicator to the client device via a network.

3. The method of claim 1 further includes identifying changes in the animal classification event compared to previously recorded animal classification events or patterns of previously recorded animal classification events.

4. The method of claim 1, wherein the three or more load sensors individually sample the load at 20 Hz to 110 Hz.

5. The method of claim 1, wherein the three or more load sensors individually sample the load at 30 Hz to 80 Hz.

6. The method of claim 1, wherein the three or more load sensors individually sample a load of up to 10 kg.

7. The method of claim 1, wherein the animal monitoring device comprises four load sensors.

8. The method of claim 7, wherein the platform has a rectangular shape and the four load sensors are each located at a different corner of the platform.

9. The method of claim 1, wherein the animal monitoring device has three load sensors and the platform is triangular in shape.

10. The method of claim 9, wherein the platform has a triangular shape and the three load sensors are each positioned at a different corner of the platform.

11. The method of claim 1, wherein the classification of the animal behavioral attributes into animal classification events is performed using normalization logic to analyze the load data.

12. An animal monitoring system, comprising an animal monitoring device, the animal monitoring device comprising: Platform, the platform being configured to carry the contained padding material thereon; Three or more load sensors associated with the platform, wherein each of the three or more load sensors is separate from each other and receives pressure input from the platform independently, wherein each of the three or more load sensors has a sampling rate in the range of 2.5 Hz to 110 Hz; and A data communicator configured to independently transmit load data from the three or more load sensors. The animal monitoring system also includes: processor; and The memory stores instructions, which, when executed by the processor, include: Receive the payload data from the data communicator. Determine whether the load data originates from the interaction between the animal and the animal monitoring device. If the load data originates from the animal interaction, then identify the animal behavioral attributes. The animal behavioral attributes are classified into animal classification events using a machine learning classifier, and The animal's movement patterns are tracked based on the load data.

13. The animal monitoring system of claim 12, wherein the processor and the memory are physically located away from the animal monitoring device and communicate with the data communicator via a network.

14. The animal monitoring system according to claim 12, wherein: The memory also stores instructions, which, when executed by the processor, include: Identify the changes in the animal classification event compared to previously recorded animal classification events or patterns of previously recorded animal classification events.

15. The animal monitoring system of claim 12, wherein the three or more load sensors have a sampling rate of 20 Hz to 110 Hz.

16. The animal monitoring system of claim 12, wherein the three or more load sensors comprise a full-bridge configuration with circular point contacts.

17. The animal monitoring system of claim 12, wherein the three or more load sensors each have a maximum load capacity of up to 10 kg.

18. The animal monitoring system of claim 12, wherein the animal monitoring device has four load sensors and the platform is rectangular in shape, and wherein each of the four load sensors is positioned at a different corner of the platform.

19. The animal monitoring system of claim 18, wherein the platform has X-axis dimension measurements of 400 mm to 600 mm and Y-axis dimension measurements of 250 mm to 450 mm.

20. The animal monitoring system of claim 12, wherein the animal monitoring device has three load sensors and the platform is triangular in shape, and wherein the three load sensors are each located at different corners of the platform.

21. The animal monitoring system according to claim 12, wherein the platform is circular in shape.

22. A method for monitoring the health status of an animal under the control of at least one processor, comprising: Load data is obtained from three or more load sensors of an animal monitoring device, the three or more load sensors being associated with a platform on which the contained bedding material is carried, wherein each of the three or more load sensors is separate from each other and receives pressure input from the platform independently of each other; If it is determined from the load data that the interaction with the contained bedding is due to the interaction between the animal and the contained bedding, then animal behavioral attributes related to the animal are identified; Classifying the animal behavioral attributes into animal classification events using a machine learning classifier includes analyzing the load data via a stage separation algorithm, wherein the animal classification events include animal excretion, and wherein the stage separation algorithm is capable of classifying multiple discrete animal excretions occurring at multiple times; and The animal's movement patterns are tracked based on the load data.

23. The method of claim 22, wherein the multiple discrete animal excretions include both urination events and defecation events.

24. The method of claim 22, wherein the multiple discrete animal excretions include a first excretion and a second excretion, and wherein the animal classification event further includes a pre-first excretion phase prior to the first excretion and a second excretion prior to the second excretion.

25. The method of claim 24, wherein the animal sorting event further includes a post-defecation phase occurring after the second defecation.

26. The method of claim 22, wherein the classification using the stage separation algorithm determines that there are no multiple discrete animal excretions during the animal interaction, which are either no animal excretions or only a single animal excretion.

27. The method of claim 22, further comprising identifying changes in the animal classification event compared to previously recorded animal classification events.

28. The method of claim 27, further comprising associating the change in the animal classification event with physical, behavioral, or mental health problems associated with the animal.

29. The method of claim 28, wherein the physical health problem is selected from feline diseases such as urinary tract diseases, kidney disease, diabetes, hyperthyroidism, spontaneous cystitis, digestive problems, or arthritis, and the mental health problem is selected from anxiety, stress, or cognitive decline.

30. The method of claim 27, wherein the change in identifying an animal classification event includes identifying a change in animal excretion.

31. The method of claim 27, further comprising generating a notification indicating the change in the animal classification event, wherein the notification is generated after a parameter associated with the device event satisfies a threshold.

32. The method of claim 22, further comprising: The animal is identified based on the load data, and The animal is distinguished from at least one other animal that interacts with the bedding material.

33. The method of claim 22, wherein classifying the animal behavioral attributes further comprises analyzing load data from the three or more load sensors based on the interaction between the animal and the contained bedding to measure the animal's weight distribution, the location of events, the duration of events, movement patterns, entry forces, exit forces, the volatility of the animal's interactions, or combinations thereof.

34. The method of claim 22, wherein classifying the animal behavioral attributes further comprises analyzing load data from the three or more load sensors to identify or measure the animal's entry into a bedding bin containing bedding material, the amount of movement of the animal in selecting a particular excretion location, the amount of time taken to select a particular excretion location, the amount of time taken to prepare the particular excretion location, the energy taken to prepare the particular excretion location, the amount of time taken to cover the excrement, the energy taken to cover the excrement, the duration of the excretion, the total duration of the device event from the animal's entry to its exit, the weight of the excrement, the animal's movement during the excretion, step / slope detection on a single load sensor during the excretion, the animal's exit from the positioned bedding bin, one or more movements or impacts involving the bedding bin, or a combination thereof.

35. The method of claim 22, wherein classifying the animal behavioral attributes further comprises analyzing load data from the three or more load sensors in both the time domain based on time-domain features and the frequency domain based on frequency-domain features, based on the interaction between the animal and the contained bedding, wherein the time-domain features include mean, median, standard deviation, range, autocorrelation, or combinations thereof, and wherein the time-domain features are created as one or more inputs to the machine learning classifier, and wherein the frequency-domain features include median, energy, power spectral density, or combinations thereof, and wherein the frequency-domain features are created as one or more inputs to the machine learning classifier.

36. The method of claim 22, wherein classifying the animal behavioral attributes further comprises analyzing load data from the three or more load sensors based on the interaction between the animal and the contained bedding to determine the animal’s movement pattern, the movement pattern including coverage distance, speed, acceleration, direction of movement, alignment, distance from the point of entry into the bedding bin positioned on the platform to the center of the bedding bin, excretion location, resting location, preferred quadrant of the bedding bin, or a combination thereof.

37. The method of claim 22 further includes generating an animal behavior model for each animal based on animal classification events unique to each animal.

38. The method of claim 22, wherein the three or more load sensors each have a sampling rate of 2.5 Hz to 110 Hz.

39. The method of claim 22, wherein the three or more load sensors each have a maximum load capacity of up to 10 kg.

40. An animal monitoring system, comprising an animal monitoring device, the animal monitoring device comprising: Platform, the platform being configured to carry the contained padding material thereon; Three or more load sensors associated with the platform, wherein each of the three or more load sensors is separate from each other and receives pressure input from the platform independently; A data communicator configured to independently transmit load data from the three or more load sensors; processor; and The memory stores instructions, which, when executed by the processor, include: Load data is obtained independently from the three or more load sensors. If the interaction with the platform is determined to be due to animal interaction based on the load data, then animal behavioral attributes related to the animal are identified. The animal behavioral attributes are classified into animal classification events using a machine learning classifier, including analyzing the load data via a stage separation algorithm, wherein the animal classification events include animal excretion, and wherein the stage separation algorithm is capable of classifying multiple discrete animal excretions occurring at multiple times. The animal's movement patterns are tracked based on the load data.

41. The animal monitoring system of claim 40, wherein the stage separation algorithm is capable of separating and classifying urination events from defecation events.

42. The animal monitoring system of claim 41, wherein the phase separation algorithm is capable of classifying the pre-excretion phase before the urination event and the second excretion phase before the defecation event.

43. The animal monitoring system of claim 41, wherein the stage separation algorithm is capable of classifying the post-excretion stage that occurs after both the urination and defecation events have occurred.

44. The animal monitoring system of claim 40, wherein the stage separation algorithm is capable of determining that there are no multiple discrete animal excretions during the animal interaction, which are either no animal excretions or only a single animal excretion.

45. A non-transitory machine-readable storage medium having thereon instructions thereon, said instructions, when executed, causing a processor to perform a method for monitoring the health status of an animal, said method comprising: Load data is obtained from three or more load sensors of an animal monitoring device, the three or more load sensors being associated with a platform on which the contained bedding material is carried, wherein each of the three or more load sensors is separate from each other and receives pressure input from the platform independently of each other; If it is determined from the load data that the interaction with the contained bedding is due to the interaction between the animal and the contained bedding, then animal behavioral attributes related to the animal are identified; Classifying the animal behavioral attributes into animal classification events using a machine learning classifier includes analyzing the load data via a stage separation algorithm, wherein the animal classification events include animal excretion, and wherein the stage separation algorithm is capable of classifying multiple discrete animal excretions occurring at multiple times; and The animal's movement patterns are tracked based on the load data.

46. ​​The non-transitory machine-readable storage medium of claim 45, wherein the stage separation algorithm is capable of separating and classifying urination events from defecation events.

47. The non-transitory machine-readable storage medium of claim 46, wherein the phase separation algorithm is capable of classifying the pre-excretion phase before the urination event and the second excretion before the defecation event.

48. The non-transitory machine-readable storage medium of claim 46, wherein the stage separation algorithm is capable of classifying post-excretion stages that occur after both the urination and defecation events have occurred.

49. The non-transient machine-readable storage medium of claim 45, wherein the stage separation algorithm is capable of determining that there are no multiple discrete animal excretions during the animal interaction, either in the form of no animal excretion or only a single animal excretion.

Images