Modular interlocking pet space composite system

Abstract

The application discloses a modular interlocking pet space composite system, comprising a plurality of pet space modules that can be combined up and / or down and / or left and / or right, and an environment sensing component, a pet identity recognition component, a pet individual parameter sensing component, a pet behavior performance sensing component, an environment adjusting component, a module controller and a central computing unit. The system associates the space environment parameters, the pet individual parameters and the pet individual performance parameters with the pet identity, establishes or calls an individual environment response model based on the historical stable state of the corresponding pet, comprehensively monitors and evaluates the current state of the pet, and controls the environment adjusting component according to the evaluation result, so as to realize the closed-loop control of the pet individual state monitoring and the space environment adjusting.

Description

Technical Field

[0001] This invention relates to the field of pet products, specifically a modular interlocking pet space composite system. Background Technology

[0002] Existing pet boarding, resting, and observation facilities typically employ single cages, multi-level cages, or modular pet spaces that can be combined vertically or horizontally to improve space utilization. With increasing levels of automation, some pet spaces also include environmental monitoring devices, pet activity monitoring devices, and ventilation or temperature control systems. However, environmental monitoring, pet status monitoring, and environmental control in existing modular pet spaces are usually relatively independent. The collected environmental data lacks a stable identity correlation with individual pets, especially when different pets rotate through different space modules, making it difficult to continuously generate environmental and status data corresponding to specific pets.

[0003] Furthermore, different pets vary in size, lifestyle, and response to environmental changes. Existing equipment typically judges status and adjusts the environment based on uniform settings or current single-item test results, making it difficult to combine the historical stable state of a specific pet with the environmental parameters of the pet's space, the pet's own status parameters, and actual behavior for correlation analysis. Therefore, when a pet's activity, rest, or other states change, the equipment struggles to make a comprehensive evaluation based on the pet's long-term state and to implement targeted environmental adjustments to the pet's space based on the evaluation results, thus affecting the individualized monitoring and environmental control capabilities of modular pet spaces. Summary of the Invention

[0004] The purpose of this invention is to provide a modular interlocking pet space composite system. The system associates spatial environmental parameters, pet individual parameters, and pet individual performance parameters with the pet's identity. Based on the corresponding pet's historical stable state, it establishes or calls an individual environmental response model to comprehensively monitor and evaluate the pet's current state. Based on the evaluation results, it controls the environmental adjustment components to achieve closed-loop control of pet individual state monitoring and spatial environment adjustment.

[0005] To achieve the above objectives, the present invention proposes the following technical solution: a modular interlocking pet space composite system, comprising multiple pet space modules that can be combined vertically and / or horizontally, an environmental sensing component, a pet identification component, a pet individual parameter sensing component, a pet behavior performance sensing component, an environmental adjustment component, a module controller, and a central computing unit; Each of the pet space modules includes a module frame, a pet-accommodating space enclosed by the module frame, a front door assembly disposed at the front of the pet-accommodating space, a side protection assembly disposed at the side of the module frame, a bottom support assembly disposed at the bottom of the pet-accommodating space, and a longitudinal interlocking assembly and / or a lateral interlocking assembly disposed between adjacent pet space modules, wherein the longitudinal interlocking assembly and / or the lateral interlocking assembly are used to restrict the relative movement between adjacent pet space modules; The environmental sensing component is installed in the pet housing space and is used to obtain the spatial environmental parameters of the space where the pet is located. The pet identification component is used to obtain the pet identification information of the pet entering the corresponding pet space module, and to establish the association between the pet identification information and the corresponding pet space module; The pet individual parameter perception component is used to obtain pet individual parameters corresponding to the pet identity information, and the pet behavior performance perception component is used to obtain pet individual performance parameters corresponding to the pet identity information; The environmental regulation component is disposed in the pet space module and is used to regulate at least one environmental parameter of the pet accommodating space; The module controller is connected to the environmental perception component, the pet identification component, the pet individual parameter perception component, the pet behavior perception component, and the environmental regulation component, respectively. The central computing unit is communicatively connected to at least one of the module controllers, and is used to associate the spatial environment parameters, pet individual parameters, and pet individual performance parameters with the corresponding pet identity information, and to establish or call an individual environmental response model based on the historical stable state data associated with the corresponding pet, to comprehensively monitor and evaluate the current pet status, and to generate control instructions for the environmental regulation component based on the monitoring and evaluation results.

[0006] Furthermore, in this invention, the longitudinal interlocking assembly includes a positioning part and a limiting part that cooperate with each other, as well as a locking part for preventing adjacent pet space modules from disengaging from each other. The lateral interlocking assembly includes interlocking guides that cooperate with each other and locking parts for maintaining the connection between adjacent left and right pet space modules. The front door assembly is closably connected to the module frame, and the side protection assembly includes a mesh protection section and / or a shielding protection section to create a relatively independent pet activity space while maintaining ventilation in the pet housing space.

[0007] Furthermore, in this invention, the environmental sensing component includes at least two of a temperature and humidity sensor, a carbon dioxide sensor, a gas sensor, and an air flow detection component, and selectively includes a noise sensor and an illuminance sensor. The gas sensor is an ammonia sensor and / or a volatile organic compound sensor; The airflow detection component includes a wind speed detection component and / or a differential pressure detection component; When the airflow detection component uses a differential pressure detection component, the two detection ends of the differential pressure detection component are respectively connected to the pet holding space and the corresponding ventilation channel to obtain the pressure difference between the pet holding space and the ventilation channel, and to characterize the air exchange state of the pet holding space. The environmental control components include at least one of a speed-regulating fan, an air inlet valve, an air outlet valve, a temperature control component, a humidity control component, and a lighting component.

[0008] Furthermore, in this invention, the pet identification component includes a wireless identification reader disposed in the pet space module and an electronic identification tag (32) carried by the pet. The pet individual parameter sensing component includes multiple pressure sensing components located at different positions below the bottom support component, a non-contact body surface temperature detection component located at the top and / or side of the pet housing space, and a non-contact micro-motion detection component. The pressure sensing component is used to obtain the pet's weight characteristics based on the total force at different detection locations, and to obtain the pet's center of gravity position and activity status based on the force distribution between different detection locations and its changes over time. The non-contact body surface temperature detection component is used to obtain the body surface temperature characteristics of the pet. The non-contact micro-motion detection component is used to acquire periodic micro-motions of the pet's body and extract respiratory state features.

[0009] Furthermore, in this invention, the pet behavior perception component includes at least two of the following: the pressure perception component, the non-contact micro-motion detection component, the food bowl weight detection component for detecting changes in food weight, and the drinking water weight detection component for detecting changes in drinking water weight. The central computing unit is configured to determine the pet's activity level, resting status, and spatial position change characteristics based on the force position changes of the multiple pressure sensing components, determine the feeding frequency and amount based on the weight change of the food bowl weight detection component, determine the drinking frequency and amount based on the weight change of the water weight detection component, and determine the pet's breathing status based on the periodic micro-pressure changes detected by the non-contact micro-motion detection component and / or the pressure sensing components when the pet is resting. The central computing unit is also configured to make a consistency judgment on different sensory data that can characterize the same pet's state, and determine the dynamic reliability of the corresponding sensory data based on the consistency judgment result.

[0010] The method for intelligent monitoring and environmental control of pet status based on the modular interlocking pet space composite system described above includes: The pet identification component obtains the pet's identity information upon entering the pet space module and associates the pet's identity information with the corresponding pet space module. The spatial environment parameters, pet individual parameters, and pet individual performance parameters associated with the pet's identity information are acquired synchronously through the environmental perception component, the pet individual parameter perception component, and the pet behavior performance perception component, and multimodal state data of the corresponding pet are formed. The central computing unit establishes or calls an individual environmental response model corresponding to the pet's identity based on the pet's historical stable state data. The individual environmental response model is used to characterize the correspondence between spatial environmental parameters, pet individual parameters, and pet individual performance parameters. Based on the historical changes of the spatial environmental parameters, the cumulative environmental exposure characteristics of the corresponding pets are determined, and the response time delay between different environmental parameters and different individual pet behaviors is determined based on the change sequence of spatial environmental parameters and the change sequence of individual pet performance parameters. Based on the spatial environment parameters after time alignment according to the response time delay, the cumulative environmental exposure characteristics, the current pet individual parameters, and the historical pet individual performance parameters, the expected individual performance of the corresponding pet under the current environmental conditions is obtained through the individual environmental response model. The actual individual pet performance is compared with the expected individual performance to obtain the degree of deviation of the corresponding individual pet performance. Based on the degree of consistency between the actual individual pet performance change and the individual performance change predicted by the current spatial environment parameters, the environmental explanatory power of the current pet state change is obtained. The current state of the corresponding pet is comprehensively monitored and evaluated based on the degree of deviation of the pet's individual parameters from its stable state, the degree of deviation of its individual performance, the cumulative environmental exposure characteristics, and the degree of environmental explanation.

[0011] Furthermore, in this invention, forming the multimodal state data includes: The spatial environment parameters, pet individual parameters, and pet individual performance parameters acquired by the environmental perception component, pet individual parameter perception component, and pet behavior performance perception component are time-aligned according to the acquisition time. The dynamic reliability of the corresponding perception data is determined based on at least one of the following: the degree of consistency between different perception results corresponding to the same pet state, the working state of the corresponding perception component, and the current activity state of the pet. Based on the dynamic credibility, the corresponding perceived data is filtered, weighted, or fused; When the dynamic reliability of a type of sensory data is lower than the preset reliability threshold, the weight of that type of sensory data in the pet status monitoring and evaluation is reduced, or the estimated value of the corresponding status parameter is obtained based on other sensory data and historical status data associated with that type of sensory data. Specifically, when the pet is in a resting state, respiratory status information is obtained based on the detection results of the non-contact micro-motion detection component and the periodic micro-pressure changes of the pressure sensing component, and the dynamic reliability of the respiratory status information is determined based on the consistency between the two.

[0012] Furthermore, in this invention, establishing the individual environmental response model includes: We obtain spatial environmental parameters, individual pet parameters, and individual pet performance parameters for multiple pets under stable conditions, and establish a population-based model to characterize the general laws of pet environmental response. Obtain historical stable state data of the target pet, and establish individual adaptation parameters and / or individual adaptation layers corresponding to the identity of the target pet based on the group basic model to form the individual environmental response model; When the amount of valid historical data of the target pet is less than the preset amount of data, the individual environmental response model obtains the prediction results mainly based on the population basic model; as the amount of valid historical data of the target pet increases, the influence of the individual adaptation parameters and / or the individual adaptation layer on the prediction results is increased. The determination of the cumulative environmental exposure characteristics includes: determining the degree of deviation of each spatial environmental parameter relative to the reference state of the corresponding pet individual within a preset historical time window; configuring time decay weights for environmental deviations exceeding the preset natural fluctuation range according to the time interval from the current moment; and obtaining the cumulative environmental exposure characteristics of the corresponding spatial environmental parameters based on the weighted environmental deviations. The determination of the response time delay includes: obtaining the degree of response correlation between changes in historical spatial environmental parameters and changes in individual pet performance under multiple candidate time delays, and determining the response time delay between the corresponding spatial environmental parameters and the corresponding individual pet performance based on the degree of response correlation.

[0013] Furthermore, in this invention, determining the degree of deviation in individual performance includes: Obtain the normal fluctuation characteristics of individual pet performance parameters when the corresponding pet is in a historical stable state, as well as the correlation between different individual pet performance parameters; Based on the normal fluctuation characteristics and the correlation, the difference between the current actual individual pet performance and the expected individual performance is standardized so that the performance parameters of pet individuals with smaller normal fluctuation amplitude under historical stable conditions have a greater deviation effect than the performance parameters of pet individuals with larger normal fluctuation amplitude when they produce the same change amplitude. The determination of the degree of environmental explanation includes: obtaining the expected individual performance changes of the pet under the current environmental change conditions based on the current spatial environmental parameters and the individual environmental response model; comparing the expected individual performance changes with the current actual individual performance changes of the pet; and determining the degree of environmental explanation based on the degree to which the current actual individual performance changes match the expected individual performance changes. The higher the degree of deviation in individual performance and the lower the degree of environmental explanation, the higher the priority for determining the pet's state as requiring human attention.

[0014] Furthermore, this invention also includes: When the deviation of the individual performance and / or the pet's comprehensive monitoring and evaluation results meet the preset environmental attribution conditions, a corresponding counterfactual environmental state is constructed for at least one current spatial environmental parameter. The counterfactual environmental state is obtained by adjusting the target spatial environmental parameter to the corresponding pet's individual reference state while keeping at least one other spatial environmental parameter unchanged. Based on the current actual environmental state and the counterfactual environmental state, the corresponding pet state prediction results are obtained through the individual environmental response model. The environmental contribution of the target space environmental parameters to the current pet state change is determined according to the difference between the two. A multi-factor counterfactual environmental state is constructed for at least two spatial environmental parameters, and the degree of joint influence between the corresponding spatial environmental parameters is determined based on the difference between the predicted changes in pet status generated by the multi-factor counterfactual environmental state and the predicted changes in pet status generated by the corresponding single-factor counterfactual environmental state. Based on the comprehensive monitoring and evaluation results, the pet's current status will be determined as one of the following: stable status, environmental adaptation status, environmentally sensitive status, continuous deviation status, or artificial attention status. When the pet's current state meets the environmental regulation conditions, at least one target environmental regulation parameter is determined based on the degree of environmental contribution and / or the degree of joint influence of different spatial environmental parameters, and the target environmental regulation parameter is adjusted by the corresponding environmental regulation component controlled by the module controller. After performing environmental adjustments, the corresponding pet's spatial environmental parameters, individual pet parameters, and individual pet performance parameters are continuously acquired. When the degree of deviation of the pet's state decreases compared to before the environmental adjustment and meets the preset improvement conditions, the environmental change, environmental adjustment operation, and pet response results are used as valid individual environmental response samples to update the corresponding pet's individual environmental response model.

[0015] Beneficial effects: The technical solution of this application has the following technical effects: This invention constructs a modular pet space using multiple pet space modules that can be combined vertically and / or horizontally. Each pet space module contains an environmental perception component, a pet identification component, a pet individual parameter perception component, and a pet behavior perception component. This allows for the synchronous acquisition of environmental parameters of the pet's space, the pet's own state parameters, and the pet's behavior parameters. Simultaneously, the pet identification component establishes a connection between a specific pet and its corresponding pet space module, ensuring that different types of perception data are continuously attributed to the individual pet. This helps prevent data confusion when a pet changes spaces or when different pets rotate using the same space.

[0016] Furthermore, the central computing unit, through the module controller, aggregates spatial environmental parameters, individual pet parameters, and individual pet performance parameters associated with the corresponding pet's identity. Based on the pet's historical stable state data, it establishes or invokes an individual environmental response model. This allows for comprehensive monitoring and evaluation of the current pet's status by combining its historical state with the overall assessment, rather than relying solely on uniform thresholds or single detection parameters. When the monitoring and evaluation results meet environmental adjustment conditions, the module controller can also control environmental adjustment components to alter the environmental state of the pet's enclosure. This creates a closed loop between individual pet status monitoring and pet space environmental adjustment, improving the adaptability of the modular pet space to different pets and the targeted nature of environmental control. Attached Figure Description

[0017] The accompanying drawings are not intended to be drawn to scale. In the drawings, each identical or nearly identical component shown in the various figures may be denoted by the same reference numeral. For clarity, not every component is labeled in each figure. Embodiments of various aspects of the invention will now be described by way of example and with reference to the accompanying drawings, wherein: Figure 1 This is a schematic diagram of the structure of the present invention; Figure 2 This is a schematic diagram of the structure of the present invention; Figure 3 This is a system block diagram of the present invention; Figure 4 This is a flowchart of the method of the present invention.

[0018] The meanings of the labels in the diagram are as follows: 1. Pet space module; 11. Module frame; 12. Pet holding space; 13. Front door assembly; 14. Side protection assembly; 15. Bottom support assembly; 16. Longitudinal interlock assembly; 17. Lateral interlock assembly; 18. Ventilation channel; 2. Environmental sensing assembly; 21. Temperature and humidity sensor; 22. Carbon dioxide sensor; 23. Gas sensor; 24. Airflow detection assembly; 25. Noise sensor; 26. Illumination sensor; 3. Pet identification assembly; 31. Wireless 32. Identification reader; 4. Electronic identification tag; 5. Pet individual parameter sensing component; 6. Pressure sensing component; 7. Non-contact body surface temperature detection component; 8. Non-contact micro-motion detection component; 9. Pet behavior perception component; 10. Food bowl weight detection component; 2. Water drinking weight detection component; 11. Environmental control component; 22. Variable speed fan; 33. Air inlet valve; 44. Air outlet valve; 55. Temperature control component; 66. Humidity control component; 7. Lighting component; 8. Module controller; 9. Central computing unit. Detailed Implementation

[0019] The embodiments of the invention are described in detail below with reference to the accompanying drawings to clearly illustrate the structure, purpose, advantages, positional relationships, and connection methods of each component. It should be noted that the directional indications (such as "front," "back," "up," and "down") involved in this embodiment are based on the posture shown in the drawings and are only used to describe the relative positional relationships and movement of the components. If the posture changes, the directional indications will be adjusted accordingly. The term "connection" includes mechanical connections and electrical connections, and can be fixed connections, detachable connections, or indirect connections through an intermediate medium. The specific meaning is understood by those skilled in the art based on the context.

[0020] Example 1: Modular Interlocking Pet Space Composite System Reference Figures 1 to 3 This embodiment provides a modular interlocking pet space composite system, which includes multiple pet space modules 1, an environmental sensing component 2, a pet identification component 3, a pet individual parameter sensing component 4, a pet behavior sensing component 5, an environmental regulation component 6, a module controller 7, and a central computing unit 8. The multiple pet space modules 1 can be combined vertically and / or horizontally according to the actual number of pets and the available space, thereby forming a composite pet space with multiple relatively independent pet accommodating spaces.

[0021] Each pet space module 1 includes a module frame 11, which constitutes the main load-bearing structure of a single pet space module 1. The module frame 11 can be made of metal profiles, metal pipes, or engineering materials with sufficient load-bearing capacity to form a rectangular frame. The interior of the module frame 11 encloses a pet-accommodating space 12, which is used for a single pet to stay, rest, eat, and drink. Using relatively independent pet-accommodating spaces 12 can reduce interference caused by direct contact between different pets when multiple pets are boarded at the same time, while providing a basis for each pet to perceive its environment and monitor its condition individually.

[0022] A front door assembly 13 is provided at the front of the pet-accommodating space 12. One side of the front door assembly 13 is rotatably connected to the module frame 11 via a hinge or pivot, while the other side is detachably connected to the module frame 11 via a door lock. This allows the pet to be prevented from leaving when closed, while facilitating pet entry and exit and enabling staff to perform cleaning, feeding, and other operations when open. The front door assembly 13 is preferably a mesh door, which allows airflow even when closed and facilitates staff observation of the pet.

[0023] Side protection components 14 are provided on the sides of the module frame 11. The side protection components 14 can be configured as mesh protection, shielding protection, or a combination of both, depending on their location. The side protection components 14 located on the outside of the composite system can employ a mesh structure to improve airflow, while those located between two adjacent pet space modules 1 can employ a partial shielding structure to create visual isolation between the pet housing spaces 12, reducing the stimulation caused by adjacent pets continuously observing each other. Thus, the side protection components 14 not only restrict pets from leaving the pet housing space 12 but also provide ventilation and spatial isolation.

[0024] The pet-accommodating space 12 has a bottom support assembly 15 at its bottom. The bottom support assembly 15 includes a support plate capable of supporting the pet and a support structure located below the support plate. The support plate may be equipped with a removable pad or anti-slip mat for easy cleaning and replacement. In addition to supporting the weight of the pet, the bottom support assembly 15 is also used to install the pressure sensing assembly 41 (described later), enabling the detection of the pet's weight, center of gravity position, and changes in activity through changes in the force applied to the bottom.

[0025] To ensure a stable composite structure from multiple pet space modules 1, a longitudinal interlocking assembly 16 is provided between adjacent pet space modules 1. The longitudinal interlocking assembly 16 includes a positioning part, a limiting part that cooperates with the positioning part, and a locking part.

[0026] In this embodiment, the positioning part can be disposed on the top of the module frame 11, and the limiting part can be disposed on the bottom of another pet space module 1 at a position corresponding to the positioning part. The positioning part can be a positioning protrusion, and the limiting part can be a positioning groove adapted to the shape of the positioning protrusion. When the upper pet space module 1 is placed on the lower pet space module 1, the positioning part enters the limiting part, thereby limiting the lateral relative position of the two pet space modules 1 and reducing lateral misalignment after stacking.

[0027] The locking mechanism can employ locking pins and holes, rotary latches, or other mechanical locking structures that prevent the upper and lower modules from separating. After positioning, the locking mechanism locks the upper and lower module frames 11, preventing the upper pet space module 1 from accidentally detaching due to collisions or pet activity. The positioning and limiting parts primarily serve to provide assembly positioning and lateral limiting, while the locking mechanism primarily serves to prevent detachment. Together, they enhance the structural stability of the multi-layer module assembly.

[0028] A lateral interlocking assembly 17 is provided between adjacent pet space modules 1. The lateral interlocking assembly 17 includes a plug-in guide and a locking part. In this embodiment, the plug-in guide can be a guide rail and guide groove that slide together, so that the two pet space modules 1 are brought closer to each other in a predetermined direction during assembly, and the plug-in guide restricts the vertical misalignment between the two modules. After reaching the predetermined plug-in position, the two module frames 11 are connected by the locking part. The locking part can be a rotary latch, a pin, or a spiral quick-locking component.

[0029] With the longitudinal interlocking component 16 and the transverse interlocking component 17, users can combine the pet space module 1 into a single-row multi-layer, double-row multi-layer, or other array form according to the space and the number of pets, without having to rebuild the entire cage.

[0030] A ventilation duct 18 is provided at the rear of the pet space module 1. The ventilation duct 18 forms an air intake or exhaust path with the pet housing space 12. Multiple pet space modules 1 can each be provided with an independent ventilation duct 18, or they can be connected to a common ventilation main. The purpose of providing the ventilation duct 18 is to provide a clear detection location for the airflow detection in the environmental sensing component 2, and at the same time enable the environmental regulation component 6 to implement air exchange in the corresponding pet housing space 12.

[0031] The environmental sensing component 2 is located inside the pet housing space 12 and includes a temperature and humidity sensor 21, a carbon dioxide sensor 22, a gas sensor 23, and an airflow detection component 24, and may further include a noise sensor 25 and an illuminance sensor 26.

[0032] The temperature and humidity sensor 21 is preferably positioned within the normal height range of the pet's residence, and is spaced away from locations that directly generate temperature and humidity changes, such as air inlets and heating components, to avoid measuring localized instantaneous temperatures that may not represent the actual environment in which the pet is located. The temperature and humidity sensor 21 outputs temperature data and relative humidity data to characterize the thermal and humidity environment of the space where the pet is located.

[0033] Carbon dioxide sensor 22 is used to detect changes in carbon dioxide concentration caused by pet respiration and insufficient ventilation. Gas sensor 23 can be an ammonia sensor and / or a volatile organic compound sensor to detect changes in gas composition caused by pet excrement, pet supplies, and stagnant air. Preferably, carbon dioxide sensor 22 and gas sensor 23 are placed at the rear of the pet enclosure 12 or on the exhaust side, so that the air after passing through the pet's activity area passes through the corresponding detection area, thereby improving the representativeness of the detected air conditions to the actual environment of the pet.

[0034] The airflow detection component 24 can be either a wind speed sensor or a differential pressure detection component. When using differential pressure detection, the first detection end of the differential pressure detection component is connected to the pet housing space 12, and the second detection end is connected to the ventilation channel 18, thereby detecting the pressure difference between the pet housing space 12 and the ventilation channel 18.

[0035] The purpose of this structure is to enable the system not only to know whether the variable speed fan 61 is on, but also to determine whether air exchange has occurred between the pet enclosure 12 and the ventilation duct 18 based on the actual pressure difference. For example, when the filter component is blocked or the ventilation duct 18 is partially blocked, even if the variable speed fan 61 continues to run, the actual pressure difference and air exchange status may still change, which can be detected by the airflow detection component 24.

[0036] The noise sensor 25 is used to acquire the ambient noise intensity and duration near the pet enclosure 12. The illuminance sensor 26 is used to acquire the brightness of the space where the pet is located, so as to correlate day and night light changes with the pet's activity and rest behavior.

[0037] The pet identification component 3 includes a wireless identification reader 31 and an electronic identification tag 32. The electronic identification tag 32 is worn by the pet or fixed to the pet collar, and the wireless identification reader 31 is located on the front of the pet space module 1, at the door, or inside the pet housing space 12.

[0038] In this embodiment, an RFID electronic identification tag 32 can be used. When a pet enters the pet enclosure 12, the wireless identification reader 31 reads the unique identifier from the electronic identification tag 32 and sends the unique identifier to the module controller 7 and the central computing unit 8, thereby establishing a binding relationship between "pet identity - current pet enclosure module 1". This binding relationship ensures that subsequently collected environmental data, individual pet parameters, and behavioral data can be attributed to the specific pet, rather than just to a fixed cage.

[0039] The pet individual parameter sensing component 4 includes multiple pressure sensing components 41, a non-contact body surface temperature detection component 42, and a non-contact micro-motion detection component 43.

[0040] Multiple pressure sensing components 41 are disposed at different positions below the bottom support component 15. For example, a weighing sensor is disposed at each of the four corners of the rectangular bottom support component 15. The four pressure sensing components 41 jointly support the support plate, and when the pet is on the support plate, each pressure sensing component 41 outputs the corresponding force value.

[0041] By adding the detection values ​​of each pressure sensing component 41, the total load exerted by the pet and the support plate on each detection position can be obtained; after deducting the initial weight of the bottom support component 15 itself, the pet's weight characteristics can be obtained.

[0042] Simultaneously, the center of gravity of the pet on the bottom bearing component 15 can be calculated based on the position coordinates of each pressure sensing component 41 and the detected load. The lateral coordinate of the pet's center of gravity can be expressed as:

[0043] The vertical coordinate of a pet's center of gravity can be represented as:

[0044] in, Indicates time The pet's lateral center of gravity position; Indicates time The pet's longitudinal center of gravity position; Indicates the first A pressure sensing component 41 at time The detected pressure value; and They represent the first The lateral and longitudinal positions of the pressure sensing component 41 on the bottom support component 15; This indicates the number of pressure sensing components 41.

[0045] Through continuous calculation and Changes in these parameters can determine whether the pet has moved within the pet enclosure 12, and further establish characteristics of activity level, duration of rest, and spatial location changes.

[0046] A non-contact surface temperature detection component 42 is located on the top and / or side of the pet housing 12, and can employ an infrared thermopile sensor or an infrared temperature array. Its detection direction is towards the area where the pet primarily resides. After the pressure sensing component 41 determines that the pet has entered a valid detection position, the central computing unit 8 selects the valid detection data from the non-contact surface temperature detection component 42 to form the pet's surface temperature characteristics. The surface temperature characteristics detected here are for trend analysis and are not equivalent to the pet's core body temperature.

[0047] The non-contact micro-motion detection component 43 can employ millimeter-wave radar and be installed on the top or side wall of the pet housing 12, ensuring that the detection area covers the pet's main resting position. When the pet is resting, breathing causes periodic micro-displacements on its body surface. The non-contact micro-motion detection component 43 collects these periodic micro-displacements and obtains the respiratory frequency and rhythm changes through bandpass filtering, spectral analysis, or peak interval analysis.

[0048] Furthermore, when the pet is lying still and its center of gravity is relatively stable, the pressure sensing component 41 will also experience slight periodic force changes due to chest and abdominal breathing. The central computing unit 8 can simultaneously extract millimeter-wave micro-motion signals and pressure micro-motion signals to cross-validate the breathing detection results.

[0049] The pet behavior perception component 5 includes the aforementioned pressure perception component 41, non-contact micro-motion detection component 43, food bowl weight detection component 51, and water weight detection component 52.

[0050] The food bowl weight detection component 51 is located below the pet food bowl. After the food bowl is placed in, the initial weight of the food bowl and food is recorded. When the weight is detected to decrease in a short period of time and then remain stable, the corresponding change can be determined as the amount of food consumed at one time. By accumulating the corresponding changes over a day or other monitoring periods, the feeding frequency, the amount of food consumed at one time, and the cumulative amount of food consumed can be obtained.

[0051] The drinking water weight detection component 52 is located below the drinking water container and acquires changes in drinking water weight in a similar manner. To reduce instantaneous errors caused by water sloshing, the drinking event can be confirmed only after the weight has been continuously and stably maintained for a preset time.

[0052] The environmental control component 6 includes at least one of the following: a speed-regulating fan 61, an air inlet valve 62, an air outlet valve 63, a temperature control component 64, a humidity control component 65, and a lighting component 66.

[0053] A variable-speed fan 61 is installed in or connected to the ventilation duct 18 to change the air exchange volume. An air inlet valve 62 is installed in the air inlet path, and an air outlet valve 63 is installed in the air outlet path. The actual ventilation state of the corresponding pet-accommodating space 12 is adjusted by changing the valve opening.

[0054] The temperature control component 64 can be a heating component or an air supply control mechanism connected to an external air conditioning system. The humidity control component 65 can be a humidifying component or a dehumidifying component. The lighting component 66 is located at the top of the pet enclosure 12 and is used to adjust the illumination level of the pet enclosure.

[0055] The aforementioned environmental control components 6 are electrically connected to the module controller 7. The module controller 7 can be an embedded controller with analog input, digital input / output, and communication functions, used to complete sensor data acquisition, actuator driving, and data communication with the central computing unit 8.

[0056] The central computing unit 8 can be located within the composite system or can be a standalone edge computing host. The central computing unit 8 communicates with multiple module controllers 7 via wired bus or wireless network, and is used to store and process spatial environment parameters, pet individual parameters, and pet individual performance parameters according to the pet's identity, and to run the individual environment response model described later.

[0057] In this embodiment, the sensor data of each pet space module 1 is first appended with a module number and a collection timestamp. Then, the central computing unit 8 binds the corresponding data to the pet's identity based on the electronic identity tag 32 information obtained by the wireless identity reader 31. Thus, even if a pet is moved from one pet space module 1 to another, its historical status data continues to belong to the same individual pet, and the historical data will not be interrupted due to the change of cage location.

[0058] The working principle of this embodiment is as follows: Multiple pet space modules 1 are combined into a stable multi-layer, multi-column pet space through longitudinal interlocking components 16 and lateral interlocking components 17. After a pet enters the pet housing space 12, the wireless identity reader 31 reads the electronic identity tag 32, enabling the central computing unit 8 to determine the specific pet corresponding to the current space. The environmental sensing component 2 continuously collects spatial environmental data such as temperature, humidity, carbon dioxide, gas, and air flow; the pet individual parameter sensing component 4 collects individual parameters such as weight, body surface temperature, and respiration; and the pet behavior sensing component 5 obtains performance data such as activity, resting, eating, and drinking. The module controller 7 transmits the data to the central computing unit 8, which associates different types of data with the pet's identity and performs a comprehensive evaluation of the current state based on the intelligent monitoring method described later. When the evaluation result indicates that environmental adjustment is needed, the central computing unit 8 generates control commands, which are then controlled by the module controller 7 to operate the speed-regulating fan 61, air inlet valve 62, air outlet valve 63, temperature regulating component 64, humidity regulating component 65, and / or lighting component 66. After adjustment, continue to acquire pet response data to form a closed loop of monitoring, analysis, adjustment, and feedback updates.

[0059] Example 2: Intelligent Monitoring and Environmental Control Method for Pet Status This embodiment provides a method for intelligent monitoring of pet status and environmental control based on the above-mentioned modular interlocking pet space composite system, including the following steps.

[0060] S1. Obtain the pet's identity information and associate the pet's identity information with the corresponding pet space module 1.

[0061] It should be noted that the purpose of this step is to solve the problem of "sensor data corresponding to cage slots rather than specific pets" in modular pet spaces. Since pets may enter different pet space modules 1 at different times, if data is stored only according to module numbers, it will be difficult to form a continuous individual historical status after a pet changes cage slots.

[0062] Specifically, after a pet enters the pet space module 1, the wireless identity reader 31 reads the electronic identity tag 32 carried by the pet to obtain a unique identity identifier. Among them, subscript Representing the Only pets.

[0063] Central computing unit 8 establishes pet identification tags The module number of the current pet space module 1 The mapping relationship between them is established, and pet identification, module number, and collection time are added to the data subsequently collected by this module.

[0064] When the wireless identity reader 31 reads a new electronic identity tag 32, or when the original pet identity leaves the corresponding pet space module 1, the central computing unit 8 updates the association between the pet and the module, while the historical data is still stored under the corresponding pet identity name.

[0065] S2. Simultaneously acquire spatial environment parameters, individual pet parameters, and individual pet performance parameters to form multimodal state data.

[0066] Specifically, for the first Only pets, at all times Establish the environment state vector:

[0067] in, Indicates the first Only pets at all times The corresponding environment state vector; Indicates temperature; Indicates relative humidity; Indicates carbon dioxide concentration; Indicates the characterization values ​​for ammonia and / or volatile organic compounds; It indicates wind speed, pressure difference, or the airflow state determined by it; Indicates environmental noise parameters; This represents the illuminance parameter.

[0068] Create individual pet parameter vectors:

[0069] in, Indicates the first Individual parameter vectors for each pet; Indicates weight characteristics; Indicates the characteristics of body surface temperature; Indicates respiratory rate; It indicates the characteristics of respiratory rhythm fluctuations within a certain time window.

[0070] Establish individual pet performance vectors:

[0071] in, Represents the individual performance vector of a pet; Indicates the degree of activity; Indicates feeding characteristics; Indicates drinking water characteristics; Indicates a state of rest or lying still; It indicates the characteristics of changes in spatial location.

[0072] The degree of activity can be represented by the cumulative displacement of the pet's center of gravity per unit time:

[0073] in, Indicates statistical time window The degree of activity within; and Representing time respectively The horizontal and vertical centroid coordinates; This indicates the length of the time window used to statistically analyze pet activity levels.

[0074] When the change in the center of gravity position is less than the preset micro-motion threshold at multiple consecutive sampling times, and the non-contact micro-motion detection component 43 detects stable periodic respiratory micro-motion, it can be determined that the pet is in a resting or quiet state.

[0075] S3. Perform time alignment and dynamic credibility processing on multi-source sensing data.

[0076] It should be noted that different sensors may have different data sampling frequencies. For example, temperature and humidity data change relatively slowly, while millimeter-wave radar and bottom pressure signals have relatively high sampling frequencies. Therefore, before inputting them into subsequent models, it is necessary to map the different data onto a unified time axis.

[0077] In this embodiment, a unified data time window is established at a preset time interval. For high-frequency signals, the mean, rate of change, frequency domain features, or other statistical features within the time window can be extracted; for low-frequency environmental parameters, the most recent effective value or interpolation result can be used, thereby forming complete multimodal data within the same time window.

[0078] Furthermore, dynamic reliability is determined for various types of sensor data.

[0079] Taking respiratory data as an example, the respiratory rate detected by millimeter-wave radar is obtained simultaneously while the pet is resting. Respiratory rate obtained by bottom pressure micro-motion detection Its consistency reliability can be expressed as:

[0080] in, Indicates time The reliability of respiratory test data; This indicates the breathing frequency extracted by millimeter-wave radar; This indicates the respiratory rate extracted from the pressure micro-motion signal; This indicates the acceptable scale of difference between the two detection methods; This represents the smallest positive number that prevents the denominator from being zero.

[0081] When the two test results are close Approaching 1; as the difference between the two results increases, decline.

[0082] It should be noted that when changes in the bottom pressure center of gravity indicate that the pet is moving significantly, even if the pressure signal has a periodic component, it is not suitable to be directly used as a basis for respiratory detection. Therefore, the central computing unit 8 reduces the reliability of pressure micro-motion respiratory data and mainly uses millimeter wave micro-motion data.

[0083] Furthermore, for general state parameters They can be merged according to their credibility:

[0084] in, Indicates the fusion parameters used in the subsequent state model; Indicates the first The dynamic reliability of real-world perceived data; Represents the actual perceived value; This represents the estimated value obtained using historical states and other related parameters.

[0085] Therefore, when the sensor data is reliable, subsequent calculations mainly use the measured data; when the sensor is obstructed, the pet is moving violently, or a local fault occurs, the impact of the abnormal data on the final state judgment is reduced.

[0086] S4. Establish or invoke the individual environmental response model for the corresponding pet.

[0087] It should be noted that this step does not use a single fixed model for all pets, but rather adopts a "group base model + individual adaptation model" approach.

[0088] First, a basic training set is formed using historical data obtained from multiple pets during periods of stable condition. Each training sample includes at least a historical environmental state sequence, a historical individual parameter sequence, a historical behavioral performance sequence, and subsequent actual behavioral performance.

[0089] In this embodiment, the population baseline model can be constructed using a gated recurrent unit (GRU). First, the environmental state vector, individual parameter vector, and pet behavior vector are normalized respectively, and then features are concatenated to form the input vector for each time step.

[0090] The input vector from multiple consecutive time steps is fed into the GRU temporal coding layer. The GRU uses update and reset gates to preserve temporal dependency information related to the pet's environmental response, and obtains the hidden state vector after processing the historical time window.

[0091] Subsequently, the hidden state vector is input into the fully connected prediction layer, which outputs the expected performance vector of the pet at the future time point or the current target time point.

[0092] Specifically, in one implementation, the central computing unit 8 maps the data acquired by each sensing component to a unified time axis at a sampling interval of 1 minute. A five-dimensional environmental parameter vector is constructed using temperature, relative humidity, carbon dioxide concentration, ammonia concentration, and airflow status; a four-dimensional individual parameter vector is constructed using weight, body surface temperature, respiratory rate, and respiratory rhythm; and a five-dimensional individual performance vector is constructed using activity level, eating characteristics, drinking characteristics, resting characteristics, and spatial location change characteristics. For each sampling moment, the five-dimensional cumulative environmental exposure features and time-delay aligned environmental parameters are obtained according to the environmental cumulative exposure calculation method and response time delay determination method described later. The time-delay aligned environmental parameters, environmental cumulative exposure features, current individual parameters, and historical individual performance parameters are sequentially concatenated to form a nineteen-dimensional model input feature. The historical individual performance parameters used for current individual performance prediction do not include the actual performance data at the corresponding prediction target moment to avoid training label leakage. A historical input sequence is constructed using data from 60 consecutive sampling moments, and the five-dimensional individual performance vector actually collected at the corresponding prediction target moment is used as the supervision label. The basic model consists of an input layer, a GRU temporal coding layer, and a fully connected prediction layer connected sequentially. The GRU temporal coding layer uses a single-layer structure with 64 hidden units. The fully connected prediction layer maps the 64-dimensional hidden states to five-dimensional individual performance prediction results. During model training, the input features of each dimension are first standardized according to the mean and standard deviation of the training data. The mean squared error between the prediction results and the supervision labels is used as the loss function. The Adam optimization algorithm is used to update the model parameters. The initial learning rate is set to 0.001, the batch size is set to 32, and the maximum number of training epochs is set to 100. The training set and validation set are divided according to individual pets. The model convergence is judged by the validation set loss. Training is stopped when a lower validation loss is not obtained after 10 consecutive epochs of training, and the model parameters at the lowest validation loss are saved. In the actual operation of the model, the same data sampling interval, feature arrangement order, and standardized parameters as in the training phase are used to process real-time data. The preprocessed nineteen-dimensional input features are fed into the trained individual environment response model to obtain the five-dimensional expected individual performance corresponding to the current pet identity and environmental conditions.

[0093] Therefore, the group basic model can be represented as:

[0094] in, Represents a group-based environmental response model; Indicates the parameters of the population basic model; This represents the historical observation window used for inputting the model; Indicates the lead time for forecasting; This indicates the predicted pet behavior.

[0095] Furthermore, for specific pets with existing historical data, an individual adaptation layer for the corresponding pet is added at the output of the group basic model or after the intermediate hidden state. ,get:

[0096] in, Indicates the first Individual adaptation layer corresponding to each pet; This represents the learnable parameters of the individual adaptation layer.

[0097] When a new pet enters the system and there is insufficient historical data for the individual, the parameters of the population base model can be frozen, and only the population base model can be used for initial prediction. As the effective stable state data of the pet gradually increases, the individual adaptation layer can be updated using the pet's own data.

[0098] This avoids the inability to monitor pets using the device for the first time due to a lack of historical data. At the same time, as the usage time increases, the model gradually adapts to the pet's own activity, drinking, resting and environmental response habits.

[0099] Furthermore, to ensure that pets entering the system for the first time can complete status monitoring normally, the central computing unit 8 pre-stores reference values ​​for environmental parameters, individual parameters, and individual performance parameters determined using historical stable state data of the group. The historical stable state data is differentiated based on pet identity relationships, prioritizing data from normal activity and rest periods when the sensing components are functioning properly, the pet's identity information is clear, no significant interference events such as cage changes or equipment cleaning have occurred, and these data have been confirmed by the pet management personnel. For each type of parameter, the median of the historical stable state data is used to determine the reference value, and the initial normal fluctuation range is formed using the 5th to 95th percentiles of the corresponding historical data. For the correlation between different individual performance parameters, a covariance matrix is ​​calculated based on the historical stable state data. When a target pet enters the system for the first time or when its effective historical data is insufficient, the group reference value is used as the initial individual reference state, and the prediction result is obtained using the group-based model. Once the pet has accumulated a preset amount of valid data covering normal activity and rest periods, the individual reference value, normal fluctuation range, and covariance matrix are recalculated based on its historical data. Simultaneously, the individual adaptation layer parameters are updated using the pet's valid data, gradually transitioning the prediction process from a population-based model to an individual environmental response model. For the dynamic reliability threshold, it can be determined based on the distribution of detection differences for the same state parameter by different sensing components under historical stable conditions, with the 95th percentile of this distribution used as the consistency judgment threshold. For the natural fluctuation range of cumulative environmental exposure, it is determined using the historical normal fluctuation range of the corresponding environmental parameter. For the state classification threshold, the initial judgment boundary is determined using the comprehensive state index distribution of historical stable state samples, combined with the duration of individual deviation. When valid historical data is insufficient to form a reliable individual reference state, the population reference state continues to be used. Pre-set safety adjustments or manual alerts are implemented for abnormal environmental states exceeding the effective training data range of the population model, avoiding excessive environmental adjustments based solely on prediction results lacking historical evidence.

[0100] S5. Calculate the cumulative exposure characteristics of different environmental factors.

[0101] It should be noted that this step is used to address the problem that simply using the current instantaneous values ​​of the environment cannot reflect the continuous effects of the environment.

[0102] For the Only pets and first place Class environment parameters, in length of Calculate cumulative exposure within the historical window:

[0103] in, Indicates the first Environmental factors cutoff time The characteristics of cumulative environmental exposure formed; Indicates time The corresponding number Class environment parameters; Indicates the first Individual reference status for each pet's corresponding environmental parameters; Indicates the historical exposure time window; This represents the time weight that changes with the distance of the data from the current time. This represents a function that converts environmental deviations into effective exposure levels.

[0104] For example:

[0105] in, This represents the difference between actual environmental parameters and individual reference conditions; Indicates the first The allowable natural fluctuation range of environmental parameters.

[0106] Therefore, when environmental parameters change only slightly within the normal range of an individual's fluctuations, they do not accumulate into significant exposure; when environmental parameters deviate continuously, the cumulative environmental exposure characteristics increase with the duration.

[0107] Time weights can be applied using exponential decay:

[0108] in, This indicates the time interval between historical data and the current moment. Indicates the first Only pets to the first Time decay parameters corresponding to environmental factors.

[0109] This setting allows more recent environmental exposures to have a greater impact on the current state, while still preserving the cumulative effect of earlier environmental changes.

[0110] S6. Determine the response time delay between different environmental parameters and different individual pet behaviors.

[0111] When pets are exposed to environmental changes, their behavior does not usually change immediately. Therefore, the central computing unit 8 analyzes multiple candidate time delays.

[0112] No. Class environment parameters and the first Between individual pet-like behaviors, in time delay The degree of correlation under certain conditions can be expressed as:

[0113] in, Indicates the first The first pet Class 1 environmental factors and the first Between class behaviors, in time delay The degree of correlation of the response; Indicates advance Environmental parameters over time; Indicates the parameters of the current behavior; This represents the correlation calculation function.

[0114] Determined within the candidate timeframe:

[0115] in, Indicates environmental parameters and behavioral performance The main response delay between; This indicates a pre-defined candidate delay range; This indicates the time delay that maximizes the absolute value of the correlation.

[0116] For example, when the correlation between a pet's temperature change and activity level is highest after a 20-minute delay, the system will use 20 minutes as the main response lag for the pet's "temperature-activity performance" in the current historical period.

[0117] Instead of simply matching environmental parameters and behavioral parameters at the same time, a time-delay aligned sequence of environmental inputs is constructed.

[0118] S7. Predict the individual behaviors that pets should exhibit under current environmental conditions.

[0119] After completing the cumulative environmental exposure and time delay analysis, the time-delay aligned environmental parameters, cumulative environmental exposure characteristics, current pet individual parameters, and historical behavioral performance are input into the individual environmental response model.

[0120] The prediction process is represented as follows:

[0121] in, This indicates that the first [unit / item] already contains individual adaptation information. Pet-only environmental response model; This represents the environmental parameters after alignment with the response time delays of each environmental factor. A vector representing the cumulative exposure characteristics of various environmental factors; This indicates the current individual parameters of the pet; This indicates the pet's behavior over a period of time. This indicates the normal expected performance predicted under the current environmental and pet-specific conditions.

[0122] The practical significance of this result is not to predict "normal pet behavior" in a uniform sense, but rather to predict "the behavior that this specific pet should exhibit under the current environmental and personal conditions, based on its existing stable state data."

[0123] S8. Calculate the degree of deviation from actual performance, the degree of deviation from individual parameters, and the degree of environmental explanation.

[0124] First, determine the degree of deviation in the individual pet's behavior.

[0125] Since different performance parameters exhibit varying degrees of daily fluctuation, this embodiment uses distance based on steady-state covariance for calculation:

[0126] in, Indicates the first The degree of deviation in the current individual behavior of a single pet; This indicates the actual individual behavior of the pet; This represents the expected individual performance predicted by the model. This represents the covariance matrix formed by different performance parameters of the pet under historical stable conditions; superscript Indicates matrix transpose; This represents the inverse of the covariance matrix.

[0127] Therefore, for a behavioral parameter that has historically varied greatly, the same change in value will result in a relatively low anomalous effect; for a behavioral parameter that has been stable for a long time in history, even small changes can be identified by the model.

[0128] The degree of deviation of individual pet parameters can be expressed as:

[0129] in, Indicates the overall deviation of individual parameters; Indicates the first Current values ​​of class individual parameters; This represents the historical mean of the corresponding parameter for the pet under stable conditions; Indicates the historical fluctuation scale of the corresponding parameter; This parameter indicates the current data reliability. This indicates the number of individual parameters involved in the evaluation.

[0130] Furthermore, determine the degree of environmental explanation.

[0131] First, input the current spatial environment parameters into the individual environmental response model to obtain the expected behavior under the current environmental change conditions:

[0132] Next, obtain the pet's baseline behavior when there are no significant changes in the current environment:

[0133] The degree of environmental explanatory power can be expressed as:

[0134] in, This indicates the degree to which the current changes in pet behavior can be explained by the environment; Indicates actual behavior; This indicates the predicted behavior under current environmental conditions; This represents the baseline behavior of an individual under stable environmental conditions; Represents the L2 norm; Represents a very small positive number; This means that the calculation result will be limited to between 0 and 1.

[0135] When the actual behavior closely resembles the predicted behavior under current environmental conditions A higher value indicates that current behavioral changes can be largely explained by environmental changes; when actual behavior deviates significantly from the environmental model's predictions, reduce.

[0136] S9. Construct a counterfactual environmental state and analyze the contribution of different environmental factors to the current changes in the pet's state.

[0137] When an individual’s performance deviates to the pre-set environmental attribution conditions or the overall state reaches the pre-set environmental attribution conditions, the central computing unit 8 does not simply judge the cause based on which environmental parameter has the highest value, but instead constructs counterfactual environments item by item.

[0138] Assume the current environment is as follows:

[0139] Regarding the first Replace the environmental parameter with the corresponding individual reference value for the pet, while keeping the other environmental parameters at their current values ​​to obtain the counterfactual environment:

[0140] in, This means only removing the first... The counterfactual environment formed after the deviation from the objective environment; This indicates the pet's number Individual reference values ​​for class-specific environmental parameters.

[0141] The risk of pet status was calculated under both real-world and counterfactual conditions, and the difference was taken as the contribution of that environmental factor:

[0142] in, Indicates the first The degree to which environmental factors contribute to changes in the current state of pets; This indicates the overall risk level of the corresponding pet in the actual environment; This indicates that only the first The state risk is predicted after the environmental parameters are restored to the individual reference state.

[0143] The larger the value, the more significant the decrease in the pet's condition risk if the environmental factor is removed, and therefore the greater the contribution of the environmental factor to the current change in the pet's condition.

[0144] Specifically, during counterfactual environment analysis, the central computing unit 8 first inputs the current actual environmental parameters into the individual environmental response model to obtain the pet's expected performance under the current environmental conditions. Using the baseline performance corresponding to the individual's historical stable state as a reference, it calculates the degree of deviation of the predicted performance under the current environmental conditions. Subsequently, for the target environmental parameter to be analyzed, this parameter is replaced with the corresponding individual environmental reference value for the pet. Other environmental parameters, pet identity, current individual parameters, and model parameters remain unchanged, and the model input is reconstructed using the same data preprocessing method to obtain the expected performance under counterfactual environmental conditions. When calculating the cumulative environmental exposure characteristics, historical environmental data remains unchanged. Only candidate environmental states are applied to environmental changes that have not yet occurred within the corresponding prediction time window, avoiding the erroneous application of hypothetical environmental adjustments to past exposure data. For the current actual environment and the counterfactual environment, the predicted state risk is calculated according to the standardized deviation of the predicted performance from the individual baseline performance, the corresponding cumulative environmental exposure characteristics, and the deviation of the individual parameters, using the same weights and normalization rules. The predicted state risk is used for counterfactual comparisons between environmental factors and is distinguished from the comprehensive state risk index calculated based on current actual perceived data. The difference between the predicted state risks under the two environmental conditions is used to determine the environmental contribution of the target environmental parameters. For multi-factor counterfactual environments, the corresponding predicted state risks are calculated in the same way, and the degree of joint impact is determined based on the difference between the sum of the changes in multi-factor risks and the changes in single-factor risks. The central computing unit 8 selects target environmental parameters that can be adjusted by the environmental adjustment components 6 according to their environmental contribution, determines the adjustment direction and target value by combining the individual reference range of the corresponding parameters, outputs control commands to the module controller 7, and drives the corresponding environmental adjustment components 6 to perform graded adjustment. After adjustment, environmental parameters and pet status data are re-collected to verify the actual adjustment effect.

[0145] Furthermore, to identify the combined effect of two environmental factors, the first one can be simultaneously included. Item and the The environmental parameters were restored to individual reference values, resulting in a two-factor counterfactual environment.

[0146] The two-factor interaction contribution can be expressed as:

[0147] in, Indicates the first Class 1 environmental factors and the first The degree of combined influence among environmental factors; This indicates the change in state risk resulting from the simultaneous removal of two environmental deviation factors. and These represent the changes in risk resulting from the removal of the corresponding environmental factor individually.

[0148] When the combined effect is significantly greater than the simple sum of the effects of two single factors, it can be determined that the pet has a significant combined response to the corresponding combination of environmental factors.

[0149] S10. Obtain the pet's overall status and classify it.

[0150] The comprehensive state index can be expressed as:

[0151] in, This indicates the pet's current overall risk index; Indicates the degree of deviation of individual pet parameters; Indicates the degree of deviation in an individual pet's behavior; This represents the total environmental exposure amount obtained by further summarizing the cumulative exposure characteristics of various environmental factors; Indicates the degree of environmental explanation; , , and These represent the weights of the corresponding evaluation items; This represents a normalization function that maps the result to a preset range.

[0152] in:

[0153] This indicates the portion of behavioral deviation that cannot be explained by the environment. When the behavioral deviation is large but the current environmental changes can adequately explain the corresponding changes, this item is relatively low; when the behavioral deviation is significant and the environmental explanation is low, this item increases.

[0154] Furthermore, based on the overall condition, the degree of environmental explanation, and the duration of individual deviation, the current pet's condition is classified into a stable condition, an environmentally adapted condition, an environmentally sensitive condition, a persistent deviation condition, or a condition requiring human intervention.

[0155] Among them, the stable state means that the spatial environment parameters, pet individual parameters, and pet individual performance are all within the historical stable range of the corresponding pet.

[0156] Environmental adaptation status indicates that a pet's individual behavior has changed to some extent, but the corresponding changes can be fully explained by recent environmental changes and the pet's existing environmental response patterns.

[0157] An environmentally sensitive state indicates that some environmental parameters may still be within the system's generally permissible range, but relative to the pet's own historical state, there has been a significant cumulative exposure or counterfactual contribution, which has already caused significant changes in individual behavior.

[0158] A sustained deviation indicates that one or more individual pet parameters or individual performance parameters continuously deviate from the individual's stable state.

[0159] The status of being under human attention indicates that the pet's individual parameters and / or behavioral performance have deviated significantly and persistently, and the current environmental changes cannot fully explain the corresponding deviations.

[0160] It should be noted that the above status is used for status monitoring and alerts on the pet space device side, and is not used to output specific disease names or medical diagnosis conclusions.

[0161] S11. Implement targeted environmental adjustments based on the degree of environmental contribution.

[0162] When a pet is in an environmentally adapted or environmentally sensitive state, and environmental contribution analysis indicates that one or more environmental factors are related to the current state change, the central computing unit 8 determines the target environmental parameters based on the degree of each environmental contribution.

[0163] For example, when ammonia and / or volatile organic compounds contribute significantly to the environment, the speed of the variable speed fan 61 can be increased and / or the opening of the air inlet valve 62 and the air outlet valve 63 can be enlarged to improve air exchange in the pet enclosure 12.

[0164] When the temperature contribution is high, the temperature of the corresponding pet-accommodating space 12 is changed by the temperature adjustment component 64.

[0165] When the humidity contribution is high, humidification or dehumidification is performed through the humidity adjustment component 65.

[0166] When temperature and humidity exhibit a significant combined effect, temperature control component 64 and humidity control component 65 can be controlled simultaneously, rather than adjusting only one of them.

[0167] To avoid sudden changes in environmental parameters causing new stress to pets, environmental regulation can be output in a tiered manner. For example, only a portion of the fan speed, valve opening, or temperature setpoint can be changed in each control cycle, and the status can be re-collected in the next cycle before deciding whether to make further adjustments.

[0168] S12. Verify the effectiveness of environmental regulation and update the individual environmental response model.

[0169] After environmental conditioning is completed, continue to acquire spatial environmental parameters, individual pet parameters, and individual pet performance parameters, and wait for the learned response time between the corresponding environmental factors and pet behavior.

[0170] The improvement in the overall state deviation before and after adjustment is defined as follows:

[0171] in, This indicates the amount of improvement in condition resulting from this environmental adjustment; This indicates the overall degree of deviation before environmental adjustments were implemented; Indicates the elapsed response time The degree of overall deviation afterwards.

[0172] when:

[0173] At that time, this environmental adjustment was determined to be an effective adjustment. Among them, This indicates the preset effective improvement threshold.

[0174] Furthermore, the environmental change, the pet's state before adjustment, the adjustment actions performed, and the actual pet response after adjustment are used to form a new effective training sample, which is then added to the pet's individual historical dataset.

[0175] To avoid significant drift in individual models due to a single accidental change, this embodiment preferably keeps the basic parameters of the population model unchanged and only updates the individual adaptation layer parameters of the corresponding pet; after accumulating a sufficient amount of new stable state data, the individual environmental response model is then periodically updated.

[0176] Through the above methods, individual models can gradually learn the sensitivity, response delay, and behavioral recovery patterns of a specific pet to different environmental factors such as temperature, humidity, and air quality.

[0177] The overall working process of this embodiment is as follows: After the pet enters the pet space module 1, it first completes identity binding. The system synchronously collects three types of data: environmental, individual, and behavioral data, and uses a dynamic credibility mechanism to reduce the impact of motion interference and abnormal sensor data. Then, it calls the individual environmental response model formed by the "group basic model + individual adaptation layer", and combines the cumulative environmental exposure and individualized response time lag to predict the normal behavior that the pet should exhibit under the current conditions. The actual performance is compared with the predicted performance to obtain behavioral deviation and individual parameter deviation, and the degree of environmental explanation is further calculated. When there is a significant deviation, counterfactual environmental analysis is used to determine the contribution of different factors such as temperature, humidity, and air quality to the current state change and the joint influence between factors. Finally, the device monitoring results of stable state, environmental adaptation state, environmental sensitive state, continuous deviation state, or artificial attention state are formed. When intervention can be carried out through environmental improvement, the central computing unit 8 controls the environmental adjustment component 6 to perform corresponding adjustments, and uses the actual pet response after adjustment to verify the effectiveness of the adjustment and update the individual environmental response model. This forms a complete closed loop of "multimodal perception - individual modeling - time lag analysis - comprehensive evaluation - environmental attribution - environmental adjustment - feedback learning".

[0178] Example 3: A Complete Operational Case of a Modular Interlocking Pet Space Composite System This embodiment takes the status monitoring and environmental control of a single pet in a pet boarding scenario as an example to illustrate the complete working process of the modular interlocking pet space composite system, from pet identification, multimodal data acquisition, individual environmental response model invocation, pet status evaluation, counterfactual environmental analysis, environmental adjustment to feedback updates.

[0179] First, multiple pet space modules 1 are combined into a multi-layered pet boarding space via vertical interlocking components 16 and horizontal interlocking components 17. Each pet space module 1 is equipped with an environmental sensing component 2, a pet identification component 3, a pet individual parameter sensing component 4, a pet behavior perception component 5, an environmental regulation component 6, and a module controller 7. Each module controller 7 is communicatively connected to the central computing unit 8. Among them, the environmental sensing component 2 includes a temperature and humidity sensor 21, a carbon dioxide sensor 22, a gas sensor 23, and an airflow detection component 24; the pet individual parameter sensing component 4 includes a pressure sensing component 41, a non-contact body surface temperature detection component 42, and a non-contact micro-motion detection component 43; the pet behavior perception component 5 obtains the corresponding pet's behavior data through the pressure sensing component 41, the non-contact micro-motion detection component 43, the food bowl weight detection component 51, and the water weight detection component 52; the environmental regulation component 6 includes a speed-regulating fan 61, an air inlet valve 62, an air outlet valve 63, a temperature regulation component 64, and a humidity regulation component 65. Each sensing component outputs detection data to the module controller 7. The module controller 7 transmits the collected data to the central computing unit 8 and drives the corresponding environmental adjustment component 6 to work according to the control instructions returned by the central computing unit 8.

[0180] When a pet enters pet space module 1, the wireless identity reader 31 reads the electronic identity tag 32 carried by the pet to obtain the pet's unique identity identifier. The module controller 7 sends the identity identifier, the current module number, and the reading time to the central computing unit 8. The central computing unit 8 establishes a correlation between the pet's identity and the current pet space module 1, and assigns subsequently collected environmental parameters, individual parameters, and individual performance parameters to the corresponding pet. When a pet moves to another pet space module 1, the central computing unit 8 updates the correlation between the pet's identity and the module number based on the new identity reading result, and continues to access the pet's corresponding historical status data. When multiple identities appear in the same module or the current pet's identity cannot be confirmed, the use of data from the corresponding time period as a valid training sample for a single pet is paused, and an identity confirmation prompt is output to the staff to reduce the risk of data confusion between different pets.

[0181] After identity binding is completed, the environmental sensing component 2 continuously acquires the temperature, relative humidity, carbon dioxide concentration, ammonia concentration, and airflow status within the pet enclosure 12. The pressure sensing component 41 acquires the pet's weight and force data at different detection locations, and determines the pet's center of gravity and activity status based on changes in force distribution; the non-contact body surface temperature detection component 42 acquires the pet's body surface temperature characteristics; the non-contact micro-motion detection component 43 acquires respiratory rate and respiratory rhythm characteristics; the food bowl weight detection component 51 and the water weight detection component 52 acquire weight changes during eating and drinking, respectively. The above sensing data is sent to the central computing unit 8 via the module controller 7, along with the pet's identity identifier, module number, and acquisition time, forming the corresponding pet's original multimodal data.

[0182] The central computing unit 8 first preprocesses the raw multimodal data, mapping data with different sampling frequencies to the same time axis according to a uniform sampling interval of 1 minute. For pressure and micro-motion signals with high change frequency, statistical features within the corresponding time window are extracted; for environmental parameters with relatively slow changes, valid detection values ​​within the time window are used to form corresponding environmental state data. When detection data is missing or abnormal, the dynamic reliability of the corresponding data is determined based on the working status of the sensing components, the pet's current activity status, and the consistency between different sensing results. For pressure micro-motion data and non-contact micro-motion data that can jointly characterize the pet's breathing state, consistency comparison is performed when the pet is in a resting state; when the pet is obviously active or the difference between the two detection results exceeds a preset threshold, the reliability of the interfered data is reduced, and the detection data with confirmed validity is used first. For parameters for which reliable measured values ​​cannot be obtained, estimated values ​​are obtained using relevant valid data and historical state data, while retaining the reliability identifier of the corresponding data to avoid directly using abnormal data for model updates.

[0183] In this embodiment, the central computing unit 8 constructs a five-dimensional environmental parameter vector from temperature, relative humidity, carbon dioxide concentration, ammonia concentration, and airflow status; a four-dimensional individual parameter vector from weight, body surface temperature, respiratory rate, and respiratory rhythm; and a five-dimensional individual performance vector from activity level, feeding characteristics, drinking characteristics, resting characteristics, and spatial position change characteristics. The spatial position change characteristic is represented by a displacement scalar determined based on changes in the pet's center of gravity, and the airflow status is represented by the corresponding differential pressure detection value. The differential pressure detection value is used to determine changes in ventilation status and is not directly equivalent to the actual air exchange volume. The central computing unit 8 further calculates the five-dimensional cumulative environmental exposure characteristics based on historical environmental data and aligns the response time delays between different environmental parameters and individual performance parameters. The time-delay-aligned five-dimensional environmental parameters, five-dimensional cumulative environmental exposure characteristics, four-dimensional individual parameters, and five-dimensional historical individual performance parameters are then sequentially concatenated to form the nineteen-dimensional model input features.

[0184] Central computing unit 8 invokes a pre-trained population baseline model. This model employs a single-layer GRU temporal coding layer and a fully connected prediction layer. The GRU temporal coding layer has 64 hidden units, and the fully connected prediction layer outputs the five-dimensional expected pet performance. During training, the same feature composition and arrangement as the actual prediction are used. Training samples are standardized, and the five-dimensional individual performance vectors collected at a specified prediction time are used as supervision labels. For conditional predictions within a specified response time range, environmental change sequences corresponding to that prediction time range from historical samples are used to construct the training input, enabling the model to learn the correspondence between environmental changes and subsequent individual performance. The actual individual performance at the target prediction time is only used for training supervision and is not used as the prediction input for that time. After training, the model parameters and standardized parameters are saved, and the same data processing method is used to construct the model input during actual runtime to maintain consistency between the training and prediction processes.

[0185] When a target pet enters the system for the first time or when there is insufficient valid historical data, the central computing unit 8 uses the group's historical stable state data to determine the initial reference state and calls the group's basic model for prediction. As the pet's valid historical data gradually increases, the central computing unit 8 selects stable state samples from data on normal activities and rest periods that are clearly identified, have valid perception data, and have been confirmed. It determines individual parameter reference values ​​based on the median of historical data and determines the individual reference intervals for each parameter based on the normal fluctuation range. When the valid sample reaches a preset data volume and covers normal activities and rest periods, the individual adaptation layer parameters are updated using the corresponding pet's valid data, allowing the individual environmental response model to gradually adapt to the pet's own behavioral patterns. When individual historical data is insufficient or the current environmental conditions exceed the model's valid training data range, the system prioritizes using preset safety adjustment rules or manual reminders, and does not directly implement significant environmental adjustments based on prediction results lacking valid historical evidence.

[0186] After obtaining valid model input, the central computing unit 8 first calculates the current pet's cumulative environmental exposure characteristics. Specifically, for each environmental parameter, valid detection data within a preset historical time window is acquired, the degree of deviation of the corresponding data from the pet's environmental reference state is determined, and time decay weights are configured according to the time interval between the historical data and the current moment to form the cumulative exposure characteristics of the corresponding environmental parameter. Simultaneously, using the pet's historical environmental change sequence and individual behavior change sequence, the main response time delays between different environmental factors and corresponding behavioral behaviors are determined, and the historical environmental data and corresponding individual behavior data are matched according to the response time delays.

[0187] For example, when the system determines, based on valid historical data, that there is a major response time delay between changes in ammonia concentration and pet activity performance, the central computing unit 8 aligns the historical ammonia concentration sequence with the activity performance sequence according to this response time delay. For other environmental factors, their corresponding response time delays are used respectively. The central computing unit 8 inputs the time-aligned environmental parameters, cumulative environmental exposure characteristics, current individual parameters, and historical individual performance parameters into the individual environmental response model to obtain the expected individual performance required for the current monitoring and assessment. It then compares the expected individual performance with the actual individual performance at the same target time to determine the degree of deviation in the corresponding pet's individual performance.

[0188] When calculating the degree of deviation in individual performance, the central computing unit 8 calculates the standardized difference between the actual and expected performance based on the normal fluctuation characteristics and correlations of different individual performance parameters under the pet's historical stable state. For the covariance matrix formed by the historical stable state data, a preset numerical stability processing method is used to avoid calculation anomalies caused by matrix invertibility. At the same time, the central computing unit 8 calculates the degree of deviation of individual parameters such as the pet's current weight, body surface temperature, and respiratory status from the historical stable state, and obtains the environmental explanatory power based on the relationship between the predicted performance under the current environmental conditions, the individual baseline performance, and the actual performance. Subsequently, based on the degree of deviation of individual parameters, the degree of deviation of individual performance, the characteristics of cumulative environmental exposure, and the environmental explanatory power, the current comprehensive state risk index is calculated, and combined with the duration of state deviation, the pet's current state is determined as a stable state, an environmentally adapted state, an environmentally sensitive state, a state of continuous deviation, or a state requiring human intervention.

[0189] When the comprehensive monitoring and evaluation results meet the preset environmental attribution conditions, the central computing unit 8 further analyzes the correlation contribution between different environmental factors and the current state change. Taking ammonia concentration as the environmental parameter to be analyzed, the central computing unit 8 first determines the ammonia reference state corresponding to the pet, the response time delay between ammonia concentration and target behavioral performance, and the adjustable range of ammonia concentration by the environmental regulation component 6. Subsequently, taking the current time as the starting time of candidate environmental regulation, a future prediction time window that can cover the response time delay is selected, and a baseline environmental scheme that maintains the current environmental regulation state and a candidate environmental scheme that adjusts the ammonia concentration towards the corresponding individual reference state are constructed respectively.

[0190] When constructing the above environmental schemes, the central computing unit 8 maintains consistency in the historical environmental data, pet identity, current individual parameters, and individual environmental response model parameters that have occurred in both schemes. For environmental changes within the future prediction time window, the baseline environmental scheme uses candidate environmental change sequences corresponding to the current environmental conditions and existing adjustment states, while the candidate environmental scheme constructs corresponding candidate environmental change sequences based on the reference range of the target environmental parameters and the adjustability of the environmental adjustment component 6. The candidate environmental change sequences are only used for model condition prediction and are not considered as actual environmental detection results that have already occurred; when a reasonable candidate environmental change process cannot be determined, automatic adjustment is not directly implemented using the corresponding prediction results, but real-time monitoring continues or preset safety adjustments are executed.

[0191] Central computing unit 8 aligns the environmental change sequences corresponding to the baseline and candidate environmental scenarios according to the response time delay of the target behavior, and uses the same prediction target time, historical observation window, individual parameters, and model parameters in both scenarios. For cumulative environmental exposure characteristics, historical environmental exposure data that has already occurred is kept unchanged, and the corresponding predicted cumulative exposure characteristics are updated only based on the candidate environmental changes within the future prediction time window for each scenario. Subsequently, the models formed by the two scenarios are fed into the individual environmental response model trained within the corresponding prediction time range to obtain two sets of expected individual behaviors corresponding to the same future prediction time. Through the above processing, the predictive impact of candidate environmental regulation on pet behavior occurs after the corresponding response time delay, rather than directly interpreting changes in current environmental parameters as changes in current behavior.

[0192] After obtaining the expected individual performance of two groups, the central computing unit 8 uses the pet's historical stable performance as a benchmark to calculate the deviation of predicted performance under the two schemes. It then combines the corresponding predicted cumulative exposure characteristics with the consistent deviation of current individual parameters to calculate the predicted state risk for both schemes according to the same prediction risk assessment rules. This predicted state risk is used to compare the expected impact of different candidate environmental schemes and is distinguished from the comprehensive state risk index calculated based on current actual perceived data. Subtracting the predicted state risk of the candidate environmental scheme from the predicted state risk of the benchmark environmental scheme yields the environmental contribution corresponding to the ammonia concentration. A positive difference indicates that the model predicts a reduction in state risk under the candidate adjustment conditions; a negative difference does not indicate that the candidate adjustment can improve the pet's condition.

[0193] For temperature, relative humidity, and other adjustable environmental parameters, the central computing unit 8 constructs a baseline environmental scenario and a single-factor candidate environmental scenario in the same manner, calculating the environmental contribution of the corresponding environmental factors. For the combined impact of two environmental factors, both factors are simultaneously set as candidate adjustment states, and the predicted state risk corresponding to the two-factor scenario is obtained according to the same future prediction target time and risk assessment rules. The risk change generated by the two-factor scenario is then compared with the sum of the risk changes of the corresponding two single-factor scenarios to obtain the degree of combined impact. The environmental contribution and combined impact degrees are used to characterize the environmental correlation impact predicted by the individual environmental response model and are not directly considered as a verified causal relationship between actual environmental factors and changes in pet status.

[0194] In this embodiment, when the environmental contribution corresponding to the ammonia concentration meets the preset environmental regulation conditions, and the target ammonia concentration is within the achievable range of the environmental regulation component 6, the central computing unit 8 determines the ammonia concentration as the target environmental regulation parameter, and generates a control command to increase the air exchange volume based on the current ammonia concentration, the corresponding pet's individual reference range, and the adjustability of the environmental regulation component 6. The central computing unit 8 sends the control command to the module controller 7 of the corresponding pet space module 1 via a communication connection. The module controller 7 increases the speed of the speed-regulating fan 61 and increases the opening of the air inlet valve 62 and / or the exhaust valve 63 according to the control command, so as to increase the air exchange volume in the pet housing space 12.

[0195] During environmental regulation, the module controller 7 adjusts the speed of the variable-speed fan 61 and the valve opening in stages according to a preset control cycle. The airflow detection component 24 and the gas sensor 23 acquire the airflow status and ammonia concentration changes during the regulation process, and feed the detection data back to the central computing unit 8. The central computing unit 8 determines whether the ammonia concentration is moving towards the target range based on the actual detection results. When the actual concentration enters the target range, the adjustment amount is stopped from increasing further; when the concentration does not improve as expected, the operating status of the environmental regulation component 6, the airflow status, and the current environmental conditions are reassessed. If necessary, further staged adjustments are taken or manual processing prompts are output to avoid continuously increasing the adjustment amount based solely on model prediction results.

[0196] After completing the ammonia concentration adjustment, the central computing unit 8 continues to collect ammonia concentration, airflow status, individual pet parameters, and individual behavior parameters. It also sets an evaluation time window based on the previously determined response time delay between the ammonia concentration and the target pet's behavior. After the corresponding response time has elapsed, the actual individual pet behavior is re-acquired, and the deviation of the adjusted overall state is calculated and compared with the deviation obtained before adjustment using the same evaluation rules. When the deviation of the adjusted overall state decreases and the improvement reaches the preset effective improvement threshold, the adjustment is deemed to have met the effective improvement conditions. If the improvement is insufficient or the state remains abnormal, monitoring continues, and other environmental factors are reassessed, or a manual attention prompt is issued.

[0197] It should be noted that when the environmental regulation target is temperature, the response time delay between temperature and the corresponding individual performance should be used to determine the effect evaluation time window; when the environmental regulation target is ammonia concentration, the response time delay between ammonia concentration and the corresponding individual performance should be used to determine the effect evaluation time window. When two or more environmental parameters are regulated simultaneously, the corresponding evaluation results should be obtained according to the response time delay of the corresponding environmental factor, and an evaluation range that can cover the response time of each target should be selected during the joint evaluation, thereby avoiding the mixing of response times between different environmental factors.

[0198] For adjustment processes that meet the conditions for effective improvement, the central computing unit 8 records the environmental state before adjustment, individual pet parameters, individual performance parameters, the actual adjustment operations performed, the environmental changes after adjustment, and the corresponding pet's actual response as an environmental response sample, along with pet identity, module number, collection time, and data reliability information. After completing the sample validity check, the environmental response sample is used to update the individual environmental response model of the corresponding pet, allowing the model to gradually acquire the pet's response change patterns to different environmental factors. During the update process, the parameters of the population base model remain unchanged; only the individual adaptation layer parameters of the corresponding pet are updated to reduce unnecessary impact of a single adjustment sample on the population base model.

[0199] For updating individual reference states and normal fluctuation ranges, the central computing unit 8 does not directly treat all effective improvement samples as stable state samples. Instead, it only includes data that meets the stable state screening criteria into the corresponding pet's stable state dataset when the pet has completed environmental adaptation, relevant environmental parameters are within the preset reference range, individual parameters and individual performance have recovered to a stable state, and the perceived data is valid. The stable state dataset is then used to update individual parameter reference values, normal fluctuation ranges, and covariance matrices, thereby avoiding the direct use of abnormal data that has not yet recovered to stability to correct the individual reference state.

[0200] When a pet is moved from its current pet space module 1 to another pet space module 1, the wireless identification reader 31 re-acquires the corresponding pet's electronic identification tag 32 information. The central computing unit 8 updates the association between the pet's identity and the module number, and continues to access the pet's existing individual historical data and individual environmental response model. Thus, even after a pet changes spaces, the continuity of individual status monitoring, environmental response analysis, and historical data can still be maintained.

[0201] Through the above operation process, the system first uses the pet identification component 3 to establish the association between the pet and the spatial module. Then, it acquires multimodal data through the environmental perception component 2, the pet individual parameter perception component 4, and the pet behavior perception component 5. The central computing unit 8 performs time alignment, credibility processing, individual environmental response prediction, and comprehensive status evaluation. When the environmental adjustment conditions are met, the system determines the target environmental parameters through counterfactual environmental analysis. The module controller 7 drives the environmental adjustment component 6 to implement hierarchical environmental adjustment. Finally, the system verifies the adjustment effect and updates the individual environmental response model through the adjusted actual environmental parameters and pet response data, thus forming a closed-loop operation process of pet individual status monitoring, environmental adjustment, and feedback learning.

[0202] While the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the invention. Those skilled in the art can make various modifications and refinements without departing from the spirit and scope of the invention. Therefore, the scope of protection of the present invention shall be determined by the claims.

Claims

1. A modular interlocking pet space composite system, characterized in that, It includes multiple pet space modules (1) that can be combined vertically and / or horizontally, an environmental perception component (2), a pet identification component (3), a pet individual parameter perception component (4), a pet behavior perception component (5), an environmental regulation component (6), a module controller (7), and a central computing unit (8). Each of the pet space modules (1) includes a module frame (11), a pet housing space (12) enclosed by the module frame (11), a front door assembly (13) disposed on the front side of the pet housing space (12), a side protection assembly (14) disposed on the side of the module frame (11), a bottom support assembly (15) disposed at the bottom of the pet housing space (12), and a longitudinal interlock assembly (16) and / or a lateral interlock assembly (17) disposed between adjacent pet space modules (1), the longitudinal interlock assembly (16) and / or the lateral interlock assembly (17) being used to restrict relative movement between adjacent pet space modules (1); The environmental sensing component (2) is installed in the pet housing space (12) and is used to obtain the spatial environmental parameters of the space where the pet is located; The pet identification component (3) is used to obtain the pet identification information of the pet entering the corresponding pet space module (1) and establish the association between the pet identification information and the corresponding pet space module (1); The pet individual parameter perception component (4) is used to obtain pet individual parameters corresponding to the pet identity information, and the pet behavior performance perception component (5) is used to obtain pet individual performance parameters corresponding to the pet identity information; The environmental adjustment component (6) is disposed on the pet space module (1) and is used to adjust at least one environmental parameter of the pet holding space (12); The module controller (7) is connected to the environmental perception component (2), the pet identification component (3), the pet individual parameter perception component (4), the pet behavior perception component (5), and the environmental regulation component (6), respectively. The central computing unit (8) is communicatively connected to at least one of the module controllers (7) and is used to associate the spatial environment parameters, pet individual parameters and pet individual performance parameters with the corresponding pet identity information, and to establish or call an individual environmental response model based on the historical stable state data associated with the corresponding pet, to comprehensively monitor and evaluate the current pet status, and to generate control instructions for the environmental regulation component (6) based on the monitoring and evaluation results.

2. The modular interlocking pet space composite system according to claim 1, characterized in that, The longitudinal interlocking assembly (16) includes a positioning part and a limiting part that cooperate with each other, as well as a locking part for preventing the upper and lower adjacent pet space modules (1) from disengaging from each other; The lateral interlocking assembly (17) includes mutually cooperating plug-in guides and locking parts for maintaining the connection state of adjacent left and right pet space modules (1); The front door assembly (13) is detachably connected to the module frame (11), and the side protection assembly (14) includes a mesh protection section and / or a shielding protection section to form a relatively independent pet activity space while maintaining ventilation of the pet housing space (12).

3. The modular interlocking pet space composite system according to claim 1, characterized in that, The environmental sensing component (2) includes at least two of the following: a temperature and humidity sensor (21), a carbon dioxide sensor (22), a gas sensor (23), and an air flow detection component (24), and optionally includes a noise sensor (25) and an illuminance sensor (26). The gas sensor (23) is an ammonia sensor and / or a volatile organic compound sensor; The airflow detection component (24) includes a wind speed detection component and / or a differential pressure detection component; When the airflow detection component (24) adopts a differential pressure detection component, the two detection ends of the differential pressure detection component are respectively connected to the pet holding space (12) and the corresponding ventilation channel (18) to obtain the pressure difference between the pet holding space (12) and the ventilation channel (18) and characterize the air exchange state of the pet holding space (12); The environmental control component (6) includes at least one of a speed-regulating fan (61), an air inlet valve (62), an air outlet valve (63), a temperature control component (64), a humidity control component (65), and a lighting component (66).

4. The modular interlocking pet space composite system according to claim 1, characterized in that, The pet identification component (3) includes a wireless identification reader (31) installed in the pet space module (1) and an electronic identification tag (32) carried by the pet. The pet individual parameter sensing component (4) includes multiple pressure sensing components (41) located at different positions below the bottom support component (15), a non-contact body surface temperature detection component (42) located at the top and / or side of the pet housing space (12), and a non-contact micro-motion detection component (43). The pressure sensing component (41) is used to obtain the pet's weight characteristics based on the total force at different detection locations, and to obtain the pet's center of gravity position and activity status based on the force distribution between different detection locations and its change over time. The non-contact body surface temperature detection component (42) is used to acquire the body surface temperature characteristics of the pet; The non-contact micro-motion detection component (43) is used to acquire periodic micro-motions of the pet's body and extract respiratory state features.

5. The modular interlocking pet space composite system according to claim 4, characterized in that, The pet behavior perception component (5) includes at least two of the following: the pressure perception component (41), the non-contact micro-motion detection component (43), the food bowl weight detection component (51) for detecting changes in food weight, and the drinking water weight detection component (52) for detecting changes in drinking water weight. The central computing unit (8) is configured to determine the pet's activity level, resting status and spatial position change characteristics based on the force position changes of the multiple pressure sensing components (41), determine the feeding frequency and amount based on the weight change of the food bowl weight detection component (51), determine the drinking frequency and amount based on the weight change of the drinking water weight detection component (52), and determine the pet's breathing status based on the periodic micro-pressure changes detected by the non-contact micro-motion detection component (43) and / or the pressure sensing components (41) in the pet's resting state. The central computing unit (8) is also configured to make a consistency judgment on different sensory data that can characterize the same pet's state, and determine the dynamic reliability of the corresponding sensory data based on the consistency judgment result.

6. A method for intelligent monitoring and environmental control of pet status based on the modular interlocking pet space composite system according to any one of claims 1-5, characterized in that, include: The pet identity recognition component (3) obtains the pet identity information that enters the pet space module (1) and associates the pet identity information with the corresponding pet space module (1); The spatial environment parameters, pet individual parameters and pet individual performance parameters associated with the pet identity information are synchronously acquired through the environmental perception component (2), the pet individual parameter perception component (4) and the pet behavior performance perception component (5), and multimodal state data of the corresponding pet are formed. The central computing unit (8) establishes or calls an individual environmental response model corresponding to the pet's identity based on the historical stable state data of the corresponding pet. The individual environmental response model is used to characterize the correspondence between spatial environmental parameters, pet individual parameters and pet individual performance parameters. Based on the historical changes of the spatial environmental parameters, the cumulative environmental exposure characteristics of the corresponding pets are determined, and the response time delay between different environmental parameters and different individual pet behaviors is determined based on the change sequence of spatial environmental parameters and the change sequence of individual pet performance parameters. Based on the spatial environment parameters after time alignment according to the response time delay, the cumulative environmental exposure characteristics, the current pet individual parameters, and the historical pet individual performance parameters, the expected individual performance of the corresponding pet under the current environmental conditions is obtained through the individual environmental response model. The actual individual pet performance is compared with the expected individual performance to obtain the degree of deviation of the corresponding individual pet performance. Based on the degree of consistency between the actual individual pet performance change and the individual performance change predicted by the current spatial environment parameters, the environmental explanatory power of the current pet state change is obtained. The current state of the corresponding pet is comprehensively monitored and evaluated based on the degree of deviation of the pet's individual parameters from its stable state, the degree of deviation of its individual performance, the cumulative environmental exposure characteristics, and the degree of environmental explanation.

7. The intelligent pet status monitoring and environmental control method according to claim 6, characterized in that, The formation of the multimodal state data includes: The spatial environment parameters, pet individual parameters, and pet individual performance parameters acquired by the environmental perception component (2), pet individual parameter perception component (4), and pet behavior performance perception component (5) are time-aligned according to the acquisition time. The dynamic reliability of the corresponding perception data is determined based on at least one of the following: the degree of consistency between different perception results corresponding to the same pet state, the working state of the corresponding perception component, and the current activity state of the pet. Based on the dynamic credibility, the corresponding perceived data is filtered, weighted, or fused; When the dynamic reliability of a type of sensory data is lower than the preset reliability threshold, the weight of that type of sensory data in the pet status monitoring and evaluation is reduced, or the estimated value of the corresponding status parameter is obtained based on other sensory data and historical status data associated with that type of sensory data. When the pet is in a resting state, respiratory status information is obtained based on the detection results of the non-contact micro-motion detection component (43) and the periodic micro-pressure changes of the pressure sensing component (41), and the dynamic reliability of the respiratory status information is determined based on the consistency between the two.

8. The intelligent pet status monitoring and environmental control method according to claim 6, characterized in that, Establishing the individual environmental response model includes: We acquire spatial environmental parameters, individual pet parameters, and individual pet performance parameters for multiple pets under stable conditions, and establish a population-based model to characterize the general laws of pet environmental response. Obtain historical stable state data of the target pet, and establish individual adaptation parameters and / or individual adaptation layers corresponding to the identity of the target pet based on the group basic model to form the individual environmental response model; When the amount of valid historical data of the target pet is less than the preset amount of data, the individual environmental response model obtains the prediction results mainly based on the population basic model; as the amount of valid historical data of the target pet increases, the influence of the individual adaptation parameters and / or the individual adaptation layer on the prediction results is increased. The determination of the cumulative environmental exposure characteristics includes: determining the degree of deviation of each spatial environmental parameter relative to the reference state of the corresponding pet individual within a preset historical time window; configuring time decay weights for environmental deviations exceeding the preset natural fluctuation range according to the time interval from the current moment; and obtaining the cumulative environmental exposure characteristics of the corresponding spatial environmental parameters based on the weighted environmental deviations. The determination of the response time delay includes: obtaining the degree of response correlation between changes in historical spatial environmental parameters and changes in individual pet performance under multiple candidate time delays, and determining the response time delay between the corresponding spatial environmental parameters and the corresponding individual pet performance based on the degree of response correlation.

9. The intelligent pet status monitoring and environmental control method according to claim 6, characterized in that, The determination of the degree of deviation in individual performance includes: Obtain the normal fluctuation characteristics of individual pet performance parameters when the corresponding pet is in a historical stable state, as well as the correlation between different individual pet performance parameters; Based on the normal fluctuation characteristics and the correlation, the difference between the current actual individual pet performance and the expected individual performance is standardized so that the performance parameters of pet individuals with smaller normal fluctuation amplitude under historical stable conditions have a greater deviation effect than the performance parameters of pet individuals with larger normal fluctuation amplitude when they produce the same change amplitude. The determination of the degree of environmental explanation includes: obtaining the expected individual performance changes of the pet under the current environmental change conditions based on the current spatial environmental parameters and the individual environmental response model; comparing the expected individual performance changes with the current actual individual performance changes of the pet; and determining the degree of environmental explanation based on the degree to which the current actual individual performance changes match the expected individual performance changes. The higher the degree of deviation in individual performance and the lower the degree of environmental explanation, the higher the priority for determining the pet's state as requiring human attention.

10. The intelligent pet status monitoring and environmental control method according to any one of claims 6-9, characterized in that, Also includes: When the deviation of the individual performance and / or the pet's comprehensive monitoring and evaluation results meet the preset environmental attribution conditions, a corresponding counterfactual environmental state is constructed for at least one current spatial environmental parameter. The counterfactual environmental state is obtained by adjusting the target spatial environmental parameter to the corresponding pet's individual reference state while keeping at least one other spatial environmental parameter unchanged. Based on the current actual environmental state and the counterfactual environmental state, the corresponding pet state prediction results are obtained through the individual environmental response model. The environmental contribution of the target space environmental parameters to the current pet state change is determined according to the difference between the two. A multi-factor counterfactual environmental state is constructed for at least two spatial environmental parameters, and the degree of joint influence between the corresponding spatial environmental parameters is determined based on the difference between the predicted changes in pet status generated by the multi-factor counterfactual environmental state and the predicted changes in pet status generated by the corresponding single-factor counterfactual environmental state. Based on the comprehensive monitoring and evaluation results, the pet's current status will be determined as one of the following: stable status, environmental adaptation status, environmentally sensitive status, continuous deviation status, or artificial attention status. When the pet's current state meets the environmental regulation conditions, at least one target environmental regulation parameter is determined based on the degree of environmental contribution and / or the degree of joint influence of different spatial environmental parameters, and the target environmental regulation parameter is adjusted by the corresponding environmental regulation component (6) through the module controller (7); After performing environmental adjustments, the corresponding pet's spatial environmental parameters, individual pet parameters, and individual pet performance parameters are continuously acquired. When the degree of deviation of the pet's state is reduced compared to before the environmental adjustment and the preset improvement conditions are met, the environmental change, environmental adjustment operation, and pet response results are used as valid individual environmental response samples to update the corresponding pet's individual environmental response model.

Images