Pet chronic disease management system and method based on behavior-physiology conjoint analysis

Abstract

The invention relates to the technical field of pet health management systems, in particular to a pet chronic disease management system and method based on behavior-physiology conjoint analysis, and the system comprises a remote monitoring module which is used for obtaining real-time behavior parameters and key physiological parameters of a pet; the early warning module is in communication connection with the remote monitoring module and is used for receiving the behavior parameters and the physiological parameters sent by the remote monitoring module; based on the behavior parameters and the physiological parameters, identifying potential abnormity and generating early warning information; the decision module is in communication connection with the early warning module and is used for receiving the early warning information sent by the early warning module; based on the early warning information, making personalized nursing schemes and feeding suggestions by using an intervention optimization algorithm based on causal inference; the user module is in communication connection with the decision-making module and is used for receiving the nursing scheme and the feeding suggestion generated by the decision-making module; nursing schemes and feeding suggestions are displayed to users, potential health problems can be found as soon as possible, and conditions are created for early intervention.

Description

Technical Field

[0001] The present invention relates to the technical field of pet health management systems, and more specifically, to a pet chronic disease management system and method based on behavior-physiology combined analysis. Background Art

[0002] As pets play an increasingly important role in human life, pet health management, especially chronic disease management, has become an important topic in the field of pet medicine. Traditional pet chronic disease management methods mainly rely on regular physical examinations and symptom observations, which have many limitations. First, regular physical examinations cannot reflect the daily changes in the health status of pets, and it is easy to miss the best time for early intervention. Second, symptom observation often relies on the owner's experience and judgment, lacking objectivity and continuity. In addition, traditional methods make it difficult to implement personalized treatment plans and fail to fully consider the individual differences of each pet.

[0003] In recent years, with the development of wearable devices and Internet of Things technology, some pet health management systems based on remote monitoring have begun to emerge. These systems are usually able to collect some physiological data of pets in real time, such as heart rate, body temperature, etc., and transmit the data to the cloud for analysis. However, the existing systems still have some obvious shortcomings. First, most systems only focus on a single or a few physiological indicators and cannot fully reflect the health status of pets. Secondly, these systems usually use simple threshold judgment methods for anomaly detection, which makes it difficult to identify complex health problems. Furthermore, the existing systems are relatively weak in data analysis and decision support, and can often only provide basic data display and simple suggestions, making it difficult to provide pet owners and veterinarians with in-depth insights and personalized management solutions.

[0004] Most importantly, existing pet health management systems generally ignore the inherent connection between behavioral data and physiological data. In fact, changes in pet behavior are often early signs of health problems, and abnormalities in physiological indicators may also lead to behavioral changes. Ignoring this connection may lead to misjudgment of the pet's health status, thus affecting management effectiveness. Summary of the invention

[0005] In response to the above problems, the present invention proposes a pet chronic disease management system and method based on behavioral-physiological joint analysis. The system achieves comprehensive, accurate and intelligent management of pet chronic diseases by integrating functions such as remote monitoring, real-time analysis, intelligent early warning and personalized decision-making.

[0006] In order to solve the above technical problems, the present invention adopts the following technical solutions:

[0007] Pet chronic disease management system based on behavioral-physiological joint analysis, including:

[0008] Remote monitoring module for:

[0009] Obtain pets' real-time behavioral parameters and key physiological parameters;

[0010] Transmitting the behavioral parameter and the physiological parameter to a cloud database via a network;

[0011] The early warning module is connected to the remote monitoring module for:

[0012] Receiving behavioral parameters and physiological parameters sent by the remote monitoring module;

[0013] Based on the behavioral parameters and the physiological parameters, identifying potential abnormalities and generating early warning information;

[0014] A decision module is connected to the early warning module for:

[0015] Receiving the warning information sent by the warning module;

[0016] Based on the early warning information, a personalized care plan and feeding recommendations are formulated using an intervention optimization algorithm based on causal inference;

[0017] A user module is connected in communication with the decision module and is used to:

[0018] Receiving the nursing plan and feeding suggestions generated by the decision module;

[0019] The care regimen and feeding suggestions are presented to the user.

[0020] Preferably, the remote monitoring module comprises:

[0021] Wearable devices are used to collect behavioral and physiological parameters of pets;

[0022] The monitoring unit is wirelessly connected to the wearable device and is used to receive data collected by the wearable device and exchange data with the cloud.

[0023] Preferably, the wearable device comprises:

[0024] Remote measurement and control chip, used for preliminary digital-to-analog conversion of collected data;

[0025] A monitoring unit, connected to the remote measurement and control chip, includes an infrared heart rate monitoring component, an infrared temperature monitor, an acceleration sensor, and a magnetic sensor, and is used to collect the pet's heart rate, body temperature, exercise intensity, and sleep pattern data;

[0026] The wireless communication module is connected to the remote measurement and control chip and is used to transmit the collected data to the monitoring unit.

[0027] Preferably, the early warning module comprises:

[0028] Alert units, which use machine learning models to analyze pet behavior and physiological data in real time to predict potential anomalies;

[0029] The automatic adjustment unit is connected to the alarm unit and is used to automatically adjust the monitoring parameters according to the prediction result of the alarm unit.

[0030] Preferably, the machine learning model built into the alarm unit is a fused deep learning model, including a long short-term memory network and a graph neural network, which is used to learn the dynamic spatiotemporal correlation between behavioral data and heart rate and body temperature.

[0031] Preferably, the decision module comprises:

[0032] A cloud database for storing data during the management of pet chronic diseases;

[0033] An information processing module, connected to the cloud database, for analyzing and processing the stored data;

[0034] The remote control module is connected to the information processing module and is used to generate control instructions according to the analysis results.

[0035] Preferably, the information processing module has a built-in machine learning model for performing spatiotemporal correlation analysis on pet behavioral parameters and physiological parameters, and predicting potential abnormalities using a causal inference algorithm.

[0036] Preferably, the remote control module includes an intervention optimization algorithm, which adopts a Bayesian causal network causal inference algorithm model to learn the common changes of behavioral and physiological parameters and the progression of chronic diseases, and dynamically optimize the treatment plan according to the individual's physiological response characteristics.

[0037] Preferably, the user module includes a remote client for pushing notifications to users and displaying the results obtained by the early warning module and the decision-making module.

[0038] The pet chronic disease management method based on behavioral-physiological combined analysis using the system includes the following steps:

[0039] S1: Collect pets’ real-time behavioral parameters and key physiological parameters;

[0040] S2: transmitting the behavioral parameter and the physiological parameter to a cloud server;

[0041] S3: Analyze the behavioral parameters and the physiological parameters in real time using a machine learning model to identify potential abnormalities;

[0042] S4: Based on the identified potential abnormalities, develop personalized care plans and feeding recommendations using intervention optimization algorithms based on causal inference;

[0043] S5: sending the nursing plan and the feeding suggestion to the user terminal for display;

[0044] Among them, the machine learning model is a fused deep learning model, including a long short-term memory network and a graph neural network, which is used to learn the dynamic spatiotemporal correlation between behavioral data and heart rate and body temperature; the intervention optimization algorithm adopts a Bayesian causal network causal inference algorithm model, which is used to learn the common changes in behavioral and physiological parameters and the progression of chronic diseases, and dynamically optimize the treatment plan based on the individual's physiological response characteristics.

[0045] The beneficial effects of the present invention are mainly reflected in the following aspects:

[0046] First, through the design of the remote monitoring module, this system realizes the comprehensive and continuous collection of pet behavior and physiological data, overcoming the defects of discontinuous and incomplete data collection in traditional methods. This comprehensive data collection provides a solid foundation for subsequent analysis and decision-making.

[0047] Secondly, the early warning module of the present invention uses advanced machine learning algorithms, especially the deep learning model that integrates LSTM and GNN, which can effectively capture the complex relationship between behavioral data and physiological data. This method greatly improves the accuracy and sensitivity of anomaly detection, allowing the system to detect potential health problems early and create conditions for early intervention.

[0048] Furthermore, the decision module of the present invention introduces an intervention optimization algorithm based on causal inference, which is a major breakthrough in the prior art. The algorithm can not only analyze the correlation between various indicators, but also infer the potential causal relationship, thereby providing a scientific basis for formulating personalized management plans. This method greatly improves the accuracy and effectiveness of chronic disease management.

[0049] In addition, the system of the present invention realizes seamless connection of data collection, analysis, early warning and decision-making through close cooperation between modules. This synergy effect not only improves the overall efficiency of the system, but also produces some unexpected positive effects. For example, the output of the decision-making module can be fed back to the early warning module to continuously optimize the early warning model; and the results of the early warning module can guide the remote monitoring module to adjust the data collection strategy, forming a closed-loop system that continuously improves itself.

[0050] From a micro perspective, the present invention has innovations in many technical details. For example, the design of the wearable device fully considers the comfort of the pet and the accuracy of data collection; the data transmission uses an efficient and secure encryption algorithm; the cloud database uses a distributed architecture suitable for processing large-scale time series data. These innovations in details together constitute an efficient and reliable technical system.

[0051] In general, the system of the present invention achieves comprehensive, accurate and intelligent management of pet chronic diseases by integrating multiple advanced technologies. It not only overcomes many limitations of existing technologies, but also achieves a qualitative leap in data analysis depth, intelligent decision support, etc. This innovation is expected to significantly improve the management effect of pet chronic diseases, extend the healthy life span of pets, and provide more powerful decision support for pet owners and veterinarians. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] Figure 1 It is the overall logic block diagram of the system of the present invention.

[0053] Figure 2 It is a logic block diagram of the remote monitoring module of the present invention.

[0054] Figure 3 It is a logic block diagram of the early warning module of the present invention.

[0055] Figure 4 It is a logic block diagram of the policy module of the present invention.

[0056] Figure 5 It is a logic block diagram of the user module of the present invention. DETAILED DESCRIPTION

[0057] like Figure 1-5 As shown, the present invention provides a pet chronic disease management system based on behavior-physiology combined analysis, including:

[0058] Remote monitoring module 1, used for:

[0059] Obtain pets' real-time behavioral parameters and key physiological parameters;

[0060] Transmitting the behavioral parameter and the physiological parameter to a cloud database via a network;

[0061] The early warning module 2 is connected to the remote monitoring module 1 for:

[0062] Receiving behavioral parameters and physiological parameters sent by the remote monitoring module 1;

[0063] Based on the behavioral parameters and the physiological parameters, identifying potential abnormalities and generating early warning information;

[0064] The decision module 3 is connected in communication with the early warning module 2 and is used to:

[0065] Receiving the warning information sent by the warning module 2;

[0066] Based on the early warning information, a personalized care plan and feeding recommendations are formulated using an intervention optimization algorithm based on causal inference;

[0067] The user module 4 is connected in communication with the decision module 3 and is used to:

[0068] Receiving the nursing plan and feeding suggestions generated by the decision module 3;

[0069] The care regimen and feeding suggestions are presented to the user.

[0070] The remote monitoring module 1 of the present invention realizes real-time and continuous collection of pet behavior and physiological data. Preferably, the behavioral parameters include but are not limited to activity level, sleep duration, number of meals, etc.; the physiological parameters include but are not limited to heart rate, body temperature, respiratory rate, etc. The selection of these parameters is based on the empirical values ​​of pet medicine research and can fully reflect the health status of the pet. For example, the heart rate range of a normal adult dog is usually between 60-140 beats / minute, and the body temperature is between 38.3-39.2°C. Any values ​​outside these ranges may indicate potential health problems.

[0071] The early warning module 2 of the present invention uses advanced machine learning algorithms to achieve real-time monitoring of the health status of pets and abnormal early warning. The module can identify minor abnormalities that may be overlooked by traditional methods, thereby achieving early prevention of diseases. For example, if the system detects that the activity of the pet decreases by more than 30% within 24 hours and the body temperature rises by more than 0.5°C, the early warning module 2 will immediately generate an early warning message. The setting of this threshold is based on the analysis of a large amount of clinical data and can effectively balance sensitivity and specificity.

[0072] Decision module 3 is the core of the present invention. It adopts an intervention optimization algorithm based on causal inference to achieve personalized chronic disease management. The core idea of ​​the algorithm is to formulate the optimal intervention strategy by analyzing the causal relationship between behavioral parameters and physiological parameters. For example, for a cat with diabetes, the algorithm may recommend an appropriate increase in exercise or adjustment of insulin dosage based on the relationship between its blood sugar level and activity level. Compared with traditional empirical treatment, this method can control the disease more accurately and reduce the risk of complications.

[0073] In one embodiment of the present invention, the remote monitoring module 1 comprises:

[0074] Wearable device 11, used to collect behavioral parameters and physiological parameters of pets;

[0075] The monitoring unit 12 is wirelessly connected to the wearable device 11 and is used to receive data collected by the wearable device 11 and exchange data with the cloud.

[0076] The design of the wearable device 11 takes full account of the comfort of the pet and the accuracy of data collection. Preferably, the wearable device 11 is made of flexible material and weighs no more than 1% of the pet's body weight to ensure that it does not affect the pet's normal activities. For example, for a cat weighing 5kg, the weight of the wearable device 11 should not exceed 50g. This design standard is based on animal behavior research and can minimize the impact of the device on the pet's behavior, thereby ensuring the authenticity of the collected data.

[0077] The monitoring unit 12 implements preliminary processing and secure transmission of data. In a preferred embodiment of the present invention, the monitoring unit 12 uses the AES-256 encryption algorithm to encrypt data, ensuring the security of data transmission. At the same time, the monitoring unit 12 also has a local storage function, which can cache up to 72 hours of data when the network is interrupted, effectively preventing data loss.

[0078] In another embodiment of the present invention, the wearable device 11 includes:

[0079] The remote measurement and control chip 111 is used to perform preliminary digital-to-analog conversion on the collected data;

[0080] The monitoring unit 112 is connected to the remote measurement and control chip 111, and includes an infrared heart rate monitoring component 1121, an infrared temperature monitor 1122, an acceleration sensor 1123 and a magnetic sensor 1124, and is used to collect the heart rate, body temperature, exercise intensity and sleep pattern data of the pet;

[0081] The wireless communication module 113 is connected to the remote measurement and control chip 111 and is used to transmit the collected data to the monitoring unit 12 .

[0082] The remote measurement and control chip 111 adopts a low-power design, with power consumption not exceeding 10mW, and can achieve continuous operation for more than 30 days. This design greatly reduces the frequency of device charging and improves user convenience. Each sensor in the monitoring unit 112 is carefully selected and calibrated to adapt to the physiological characteristics of pets. For example, the infrared heart rate monitoring component 1121 adopts reflective photoplethysmography (PPG), which can penetrate the pet's fur to accurately measure the heart rate with a measurement accuracy of ±2 times / minute.

[0083] The wireless communication module 113 uses Bluetooth 5.0 technology with a transmission distance of up to 100 meters, which meets the needs of most home environments. At the same time, it also supports mesh network technology, which can form a network through multiple devices to expand the coverage range, and is suitable for scenarios such as large kennels or pet hospitals.

[0084] The method of the present invention achieves comprehensive and accurate management of pet chronic diseases through behavioral-physiological combined analysis. Compared with traditional methods, the present invention has the following significant advantages:

[0085] 1. Real-time: Through the cooperation of the wearable device 11 and the monitoring unit 12, 24-hour uninterrupted monitoring of the health status of the pet is achieved, and potential health problems can be discovered in time.

[0086] 2. Accuracy: The intervention optimization algorithm based on causal inference can deeply analyze the relationship between behavioral and physiological parameters and formulate more accurate treatment plans.

[0087] 3. Personalization: The system will dynamically adjust the management plan based on the individual characteristics and changes in the condition of each pet to achieve truly personalized medical care.

[0088] 4. Preventiveness: By analyzing large amounts of data, the system can predict potential health risks and achieve early prevention of diseases.

[0089] 5. Convenience: Users can check their pets’ health status and management suggestions at any time through mobile devices, which greatly improves the convenience of chronic disease management.

[0090] In summary, the pet chronic disease management system and method based on behavioral-physiological combined analysis of the present invention provides a new solution for the management of pet chronic diseases through advanced sensing technology, data analysis algorithms and artificial intelligence technology, and is expected to significantly improve the quality of life and lifespan of pets.

[0091] The early warning module 2 of the present invention comprises:

[0092] an alarm unit 21 for analyzing pet behavior and physiological data in real time using a machine learning model to predict potential anomalies;

[0093] The automatic adjustment unit 22 is connected to the alarm unit 21 and is used to automatically adjust the monitoring parameters according to the prediction result of the alarm unit 21.

[0094] In another embodiment of the present invention, the alarm unit 21 is a key component of the system of the present invention, which uses advanced machine learning models to analyze the behavior and physiological data of the pet in real time. The core task of this unit is to identify potential health anomalies in a timely manner, thereby achieving early warning. Preferably, the alarm unit 21 uses an anomaly detection algorithm based on deep learning, which can learn the behavior and physiological parameter patterns of the pet in a normal state and issue an alarm when the real-time data deviates from these patterns.

[0095] For example, for an elderly dog ​​with heart disease, the alarm unit 21 may pay special attention to the relationship between its heart rate and activity level. If the system detects an abnormal increase in heart rate without a corresponding increase in activity level, an alarm may be triggered. This method is more accurate than simple threshold judgment and can effectively reduce the false alarm rate.

[0096] The automatic adjustment unit 22 is designed to optimize the monitoring efficiency of the system. Based on the prediction results of the alarm unit 21, the automatic adjustment unit 22 can dynamically adjust various monitoring parameters, such as data collection frequency, sensor sensitivity, etc. For example, if the system detects an abnormal fluctuation in a certain physiological indicator of the pet, the automatic adjustment unit 22 may increase the sampling frequency of the indicator from once an hour to once every 10 minutes to obtain more detailed data. This adaptive mechanism not only improves the accuracy of anomaly detection, but also optimizes the energy efficiency of the system.

[0097] In another embodiment of the present invention, the machine learning model built into the alarm unit 21 of the present invention is a fused deep learning model, including a long short-term memory network (LSTM) and a graph neural network (GNN), which is used to learn the dynamic spatiotemporal correlation between behavioral data and heart rate and body temperature.

[0098] This fused deep learning model is a major innovation of the present invention. Long short-term memory networks (LSTM) are good at processing time series data and can capture the changing patterns of pet behavior and physiological parameters over time. Graph neural networks (GNNs) are good at processing complex relational data and can effectively model the mutual influence between different parameters.

[0099] In a preferred embodiment of the present invention, the structure of the LSTM network is as follows:

[0100] f t =σ(W f ·[h t-1 ,x t ]+b f )

[0101] i t =σ(W i ·[h t-1 ,x t]+b i )

[0102]

[0103]

[0104] o t =σ(W o ·[h t-1 ,x t ]+b o )

[0105] h t =o t *tanh(C t )

[0106] Among them, f t ,i t ,o t They are forget gate, input gate and output gate respectively, C t is the cell state, h t is the hidden state, x t is the input data, W and b are the weight and bias parameters.

[0107] The core operation of GNN can be expressed as:

[0108]

[0109] in, represents the feature of node v at the kth layer, N(v) represents the neighbor set of node v, and W (k) and b (k) are the weight and bias parameters of the kth layer.

[0110] The advantage of this fusion model is that it can simultaneously capture the temporal dynamics of pet behavioral and physiological parameters and the complex associations between different parameters. For example, for a cat with asthma, the model may learn the specific relationship pattern between breathing rate and activity level when the temperature changes, thereby more accurately predicting possible asthma attacks.

[0111] In a preferred embodiment of the present invention, the decision module 3 of the present invention comprises:

[0112] A cloud database 31, used to store data in the process of pet chronic disease management;

[0113] An information processing module 32, connected to the cloud database 31, for analyzing and processing the stored data;

[0114] The remote control module 33 is connected to the information processing module 32 and is used to generate control instructions according to the analysis results.

[0115] The cloud database 31 adopts a distributed storage architecture to ensure high availability and scalability of data. Preferably, the present invention uses Apache Cassandra as the database system. This choice is based on Cassandra's excellent write performance and horizontal expansion capabilities, which is very suitable for processing the large amount of time series data generated by this system.

[0116] The information processing module 32 is the core of the decision module 3, and is responsible for in-depth analysis of the massive data stored in the cloud database 31. In one embodiment of the present invention, the information processing module 32 uses the distributed computing framework Apache Spark, which can efficiently process large-scale data. This module not only performs conventional statistical analysis, but also uses machine learning algorithms to mine deep patterns in the data to provide support for personalized decision-making.

[0117] The remote control module 33 generates specific control instructions based on the analysis results of the information processing module 32. These instructions may include adjusting the feeding plan, changing the exercise recommendations, updating the medication reminder, etc. For example, for an obese dog, if the system detects that its weight continues to increase but the amount of activity is insufficient, the remote control module 33 may generate recommendations to increase exercise time and reduce food intake.

[0118] In a preferred embodiment of the present invention, the information processing module 32 of the present invention has a built-in machine learning model for performing spatiotemporal correlation analysis on pet behavior parameters and physiological parameters, and predicting potential abnormalities using a causal inference algorithm.

[0119] This built-in machine learning model is the key to the intelligence of the system of the present invention. It can not only identify the correlation between parameters, but more importantly, it can infer the potential causal relationship. The present invention adopts a causal inference algorithm based on the structural equation model (SEM), and its core idea can be expressed by the following mathematical expression:

[0120] Y=f(X,U)

[0121] Among them, Y represents the outcome variable of interest (such as a chronic disease symptom), X represents the observable covariates (such as behavioral and physiological parameters), and U represents the unobserved confounders. The function f describes how X and U affect Y.

[0122] In practical applications, the system of the present invention uses machine learning methods (such as random forests or gradient boosting trees) based on a large amount of historical data to estimate the function f. Then, by performing intervention analysis on the estimated f, the system can predict the potential impact of different behavioral or physiological parameter changes on chronic disease conditions.

[0123] For example, in a cat with diabetes, the system might learn a causal relationship between increased water intake and elevated blood sugar levels. If an abnormal increase in water intake is detected, even if blood sugar levels are within normal range, the system might alert the cat of a potential risk of losing blood sugar control.

[0124] This causal inference-based method is more explanatory and predictive than traditional correlation analysis, and can provide more reliable decision support for the early prevention and personalized management of pet chronic diseases.

[0125] In a preferred embodiment of the present invention, the remote control module 33 of the present invention includes an intervention optimization algorithm, which adopts a Bayesian causal network causal inference algorithm model to learn the common changes of behavioral and physiological parameters and the progression of chronic diseases, and dynamically optimize the treatment plan according to the individual's physiological response characteristics.

[0126] The intervention optimization algorithm of the present invention is an important innovation. It combines Bayesian causal networks with dynamic programming to achieve intelligent optimization of pet chronic disease treatment plans. The core idea of ​​the algorithm is to establish a dynamic causal model, continuously learn and update the causal relationship between various parameters, and thus predict the potential effects of different intervention measures.

[0127] In a preferred embodiment of the present invention, the Bayesian causal network can be represented as a directed acyclic graph (DAG):

[0128] G=(V,E)

[0129] Among them, V is a set of nodes, representing various behavioral and physiological parameters and symptoms of chronic diseases, and E is a set of edges, representing the causal relationship between nodes.

[0130] Each node v i ∈V has a conditional probability distribution:

[0131] P(v i |pa(v i ))

[0132] Among them, pa(v i ) indicates v i The parent node collection of .

[0133] The goal of the intervention optimization algorithm is to find the optimal intervention strategy π * , resulting in the greatest improvement in expected health status:

[0134]

[0135] Where H represents the health status indicator and do(π) represents the implementation of the intervention strategy π.

[0136] The advantage of this approach is that it can dynamically adjust treatment plans based on each pet's individual characteristics and response patterns. For example, for an elderly dog ​​with arthritis, the system may learn the complex relationship between exercise, weight, and pain levels. If it is found that increasing exercise leads to increased pain, the algorithm may recommend controlling weight first and then gradually increasing low-intensity exercise.

[0137] In one embodiment of the present invention, the user module 4 of the present invention includes a remote client for pushing notifications to users and displaying the results obtained by the early warning module 2 and the decision module 3.

[0138] The remote client is an important interface for the system of the present invention to interact with the user. Preferably, the client is in the form of a mobile application (APP) that supports both iOS and Android platforms, ensuring that most users can use it conveniently.

[0139] In one embodiment of the present invention, the main functions of the remote client include:

[0140] 1. Real-time data display: Display the pet's physiological indicators and behavioral data in an intuitive chart form. For example, a line chart can be used to show 24-hour heart rate changes, and a bar chart can be used to show daily activity.

[0141] 2. Abnormal warning reminder: When the system detects a potential health problem, the client will immediately push a notification to the user. The urgency of the notification will be graded according to the severity of the abnormality, using different prompt sounds and icons.

[0142] 3. Health management suggestions: Displays personalized care plans and feeding suggestions generated by decision module 3. These suggestions are presented in the form of a task list, and users can mark the completion status directly on the APP.

[0143] 4. Historical data review: allows users to view the pet’s historical health data and management records, helping users better understand the pet’s health trends.

[0144] 5. Communicate with veterinarians: Provide a one-click sharing function that allows users to send their pet’s health data directly to designated veterinarians, facilitating remote diagnosis and consultation.

[0145] Through these functions, the remote client of the present invention effectively converts complex data analysis results into information that is easy for users to understand and execute, greatly improving the convenience and effectiveness of pet chronic disease management.

[0146] Finally, the present invention also proposes a pet chronic disease management method based on behavioral-physiological joint analysis, comprising the following steps:

[0147] S1: Collect pets’ real-time behavioral parameters and key physiological parameters;

[0148] S2: transmitting the behavioral parameter and the physiological parameter to a cloud server;

[0149] S3: Analyze the behavioral parameters and the physiological parameters in real time using a machine learning model to identify potential abnormalities;

[0150] S4: Based on the identified potential abnormalities, develop personalized care plans and feeding recommendations using intervention optimization algorithms based on causal inference;

[0151] S5: Send the care plan and feeding suggestions to the user terminal for display.

[0152] This method achieves full-process intelligent management of pet chronic diseases by systematically integrating data collection, transmission, analysis and decision-making. In step S1, the present invention uses high-precision wearable devices to ensure the accuracy and comprehensiveness of data collection. For example, for heart rate monitoring, a PPG sensor with a sampling rate of 250Hz is used, which can accurately capture the heart rate changes of pets.

[0153] In step S3, the machine learning model of the present invention adopts the fusion deep learning model mentioned above, including LSTM and GNN. This model can effectively process time series data and complex relationship data, and is particularly suitable for analyzing the behavior-physiology joint data of pets. For example, for a cat with heart disease, the model may learn the time required for the heart rate to return to normal levels after strenuous exercise. If this time is abnormally prolonged, it may indicate a decline in heart function.

[0154] The intervention optimization algorithm in step S4 is the core innovation of this method. The algorithm not only considers the changes of a single parameter, but also the interaction between multiple parameters. For example, for a dog with diabetes, the algorithm will consider multiple factors such as blood sugar level, weight and age when making exercise recommendations, thereby generating a safer and more effective personalized plan.

[0155] In general, the method of the present invention realizes precise, personalized and intelligent management of pet chronic diseases through behavioral-physiological combined analysis, providing strong technical support for improving the quality of life and prolonging the lifespan of pets.

[0156] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto; any person familiar with the art who, within the scope disclosed by the present invention, makes substitutions or changes based on the scheme and improved concepts of the present invention shall be covered within the protection scope of the present invention.

Claims

1. A pet chronic disease management system based on behavioral-physiological joint analysis, characterized in that: include: Remote monitoring module for: Obtain pets' real-time behavioral parameters and key physiological parameters; Transmitting the behavioral parameter and the physiological parameter to a cloud database via a network; The early warning module is connected to the remote monitoring module for: Receiving behavioral parameters and physiological parameters sent by the remote monitoring module; Based on the behavioral parameters and the physiological parameters, identifying potential abnormalities and generating early warning information; A decision module is connected to the early warning module for: Receiving the warning information sent by the warning module; Based on the early warning information, a personalized care plan and feeding recommendations are formulated using an intervention optimization algorithm based on causal inference; A user module is connected in communication with the decision module and is used to: Receiving the nursing plan and feeding suggestions generated by the decision module; The care regimen and feeding suggestions are presented to the user.

2. The system according to claim 1, characterized in that The remote monitoring module comprises: Wearable devices are used to collect behavioral and physiological parameters of pets; The monitoring unit is wirelessly connected to the wearable device and is used to receive data collected by the wearable device and exchange data with the cloud.

3. The system according to claim 2, characterized in that The wearable device comprises: Remote measurement and control chip, used for preliminary digital-to-analog conversion of collected data; A monitoring unit, connected to the remote measurement and control chip, includes an infrared heart rate monitoring component, an infrared temperature monitor, an acceleration sensor, and a magnetic sensor, and is used to collect the pet's heart rate, body temperature, exercise intensity, and sleep pattern data; The wireless communication module is connected to the remote measurement and control chip and is used to transmit the collected data to the monitoring unit.

4. The system according to claim 1, characterized in that The early warning module comprises: Alert units, which use machine learning models to analyze pet behavior and physiological data in real time to predict potential anomalies; The automatic adjustment unit is connected to the alarm unit and is used to automatically adjust the monitoring parameters according to the prediction result of the alarm unit.

5. The system according to claim 4, characterized in that The machine learning model built into the alarm unit is a fused deep learning model, including a long short-term memory network and a graph neural network, which is used to learn the dynamic spatiotemporal correlation between behavioral data and heart rate and body temperature.

6. The system according to claim 1, characterized in that The decision module comprises: A cloud database for storing data during the management of pet chronic diseases; An information processing module, connected to the cloud database, for analyzing and processing the stored data; The remote control module is connected to the information processing module and is used to generate control instructions according to the analysis results.

7. The system according to claim 6, characterized in that The information processing module has a built-in machine learning model for performing spatiotemporal correlation analysis on pet behavior parameters and physiological parameters, and predicting potential abnormalities using a causal inference algorithm.

8. The system according to claim 6, characterized in that The remote control module includes an intervention optimization algorithm, which uses a Bayesian causal network causal inference algorithm model to learn the common changes in behavioral and physiological parameters and the progression of chronic diseases, and dynamically optimize the treatment plan based on the individual's physiological response characteristics.

9. The system according to claim 1, characterized in that The user module includes a remote client for pushing notifications to users and displaying the results obtained by the early warning module and the decision-making module.

10. A method for managing pet chronic diseases based on behavioral-physiological combined analysis using the system of any one of claims 1 to 9, characterized in that: The following steps are involved: S1: Collect pets’ real-time behavioral parameters and key physiological parameters; S2: transmitting the behavioral parameter and the physiological parameter to a cloud server; S3: Analyze the behavioral parameters and the physiological parameters in real time using a machine learning model to identify potential abnormalities; S4: Based on the identified potential abnormalities, develop personalized care plans and feeding recommendations using intervention optimization algorithms based on causal inference; S5: sending the nursing plan and the feeding suggestion to the user terminal for display; Among them, the machine learning model is a fused deep learning model, including a long short-term memory network and a graph neural network, which is used to learn the dynamic spatiotemporal correlation between behavioral data and heart rate and body temperature; the intervention optimization algorithm adopts a Bayesian causal network causal inference algorithm model, which is used to learn the common changes in behavioral and physiological parameters and the progression of chronic diseases, and dynamically optimize the treatment plan based on the individual's physiological response characteristics.

Images