A method and system for scene-based coordinated regulation of home environment parameters
By integrating millimeter-wave radar with environmental sensors through multimodal fusion and physiological feedback loops, the privacy risks and inter-device conflicts of home environment control devices are resolved, enabling precise scene determination and coordinated adjustment of the home environment to maximize comfort.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- QIERLING BEIJING HEALTH TECH CO LTD
- Filing Date
- 2026-03-11
- Publication Date
- 2026-06-09
AI Technical Summary
Existing home environment control devices lack a precise understanding of users' actual conditions and usage scenarios, leading to privacy risks and inter-device interference issues in multi-device linkage control.
By employing multimodal fusion of millimeter-wave radar and environmental sensors, and through scene intent determination model and environmental parameter collaborative coupling model, it achieves accurate determination of the home environment and collaborative control of multiple devices, and optimizes and adjusts by combining physiological feedback closed loop.
It achieves enhanced privacy and security without visual information, accurate scene determination, reduced false triggers, resolution of conflicts between multiple devices, maximized comfort, and automatic correction through physiological feedback, reducing the frequency of manual operation.
Smart Images
Figure CN122172608A_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the technical field of smart home and environmental perception control, and particularly relates to a method and system for scene-based collaborative adjustment of home environmental parameters. Background Art
[0002] With the popularization of smart home devices, home environmental adjustment devices are gradually developing from independent control of single devices to multi-device linkage control. In the prior art, air purifiers, humidifiers, aroma devices, lighting devices, etc. are mostly controlled by preset rules or simple sensor trigger methods, lacking an accurate understanding of the user's true state and usage scenarios.
[0003] Some solutions use cameras to obtain user behavior information, but there are problems of privacy risks and installation limitations; there are also solutions that introduce millimeter-wave radars for human presence or breathing detection, but mostly stay at the level of single perception or single-device linkage, and fail to solve the problems of mutual influence and collaborative control between multiple environmental parameters (for example: the operation of humidification devices may affect the perceived temperature, resulting in changes in temperature adjustment requirements).
[0004] Therefore, there is an urgent need for a home environmental adjustment method that does not require visual information, can comprehensively consider the user's physiological micro-motions and environmental states, and has an adaptive closed-loop optimization ability. Summary of the Invention
[0005] The purpose of the present invention is to provide a method and system for scene-based collaborative adjustment of home environmental parameters, which can accurately determine the home scene through multi-modal fusion of millimeter-wave radars and environmental sensors, and perform linkage control on multiple environmental adjustment devices based on a collaborative coupling model, and use a physiological feedback closed-loop to improve environmental comfort and reduce frequent manual intervention by users.
[0006] The present invention provides a method for scene-based collaborative adjustment of home environmental parameters, including the following steps: Step 1, collect micro-motion signals in the home space through at least one millimeter-wave radar, perform phase demodulation and time-series filtering on the micro-motion signals, and extract the human respiratory rhythm feature sequence; Step 2, synchronously collect at least one environmental parameter in the home environment, where the environmental parameter includes one or more of temperature, humidity, air quality parameter, and light intensity, and construct an environmental state vector; Step 3, perform multi-modal fusion on the respiratory rhythm feature sequence and the environmental state vector, input it into the scene intention determination model, and output the scene label corresponding to the current home environment; Step 4: Based on the scene label, call the environmental parameter collaborative coupling model to generate multi-device control action vectors, wherein the collaborative coupling model is configured to calculate the operating parameters of each environmental adjustment device based on preset device interaction constraint rules or multi-objective optimization functions. Step 5: Based on the multi-device control action vector, perform coordinated control on at least two types of environmental control devices; Step 6: Collect user's manual intervention behavior on the environmental regulation equipment and the human physiological feedback indicators after regulation, and update the parameters of the environmental parameter collaborative coupling model based on the manual intervention behavior or physiological feedback indicators to form a closed-loop optimization regulation.
[0007] Furthermore, the respiratory rhythm characteristic sequence includes at least respiratory rate, respiratory amplitude stability, and micro-motion phase continuity indicators; the artificial intervention behavior includes turning on, off, or adjusting the parameters of any environmental regulation device under automatic adjustment; the physiological feedback indicators include respiratory rate change rate, body micro-tremor signals, and body movement amplitude.
[0008] Furthermore, the millimeter-wave radar operates in the frequency band between 24 GHz and 77 GHz.
[0009] Furthermore, the multimodal fusion in step 3 includes inputting the scene intent determination model after feature concatenation or weighted fusion of the respiratory rhythm feature sequence and the environmental state vector.
[0010] Furthermore, the scene intent determination model described in step 3 is a supervised machine learning model or a deep learning model.
[0011] Furthermore, the construction method of the environmental parameter cooperative coupling model in step 4 includes any of the following: (1) Linear weighted mapping: represented as Where A is the device control action vector, S is the scene state vector, and M is the environmental parameter cooperative coupling matrix. The element weights in the matrix are used to characterize the gain or suppression relationship between different devices. (2) Nonlinear mapping: Using a fuzzy logic controller or artificial neural network, the scene labels and environmental state vectors are mapped to the device control parameters.
[0012] Furthermore, when using an environmental parameter co-coupling matrix, the weights are incrementally updated based on the frequency, direction, or magnitude of user intervention.
[0013] Furthermore, the environmental control equipment includes two or more of the following: air purification equipment, humidification equipment, fragrance release equipment, and lighting equipment.
[0014] This invention also provides a scenario-based collaborative adjustment system for home environmental parameters, comprising: Millimeter-wave sensing modules are used to collect micro-motion signals of the human body within the home space; The environmental parameter acquisition module is used to collect temperature, humidity, air quality, and light parameters; The scene intent determination module is used to perform multimodal fusion of human respiratory rhythm characteristics and environmental state and output scene labels; The collaborative control decision module is used to generate multi-device control action vectors based on a collaborative coupling model of environmental parameters. The feedback learning module is used to update the environmental parameter co-coupling model based on user manual intervention behavior and human physiological feedback indicators; Each module is configured to execute the method described.
[0015] The present invention also provides a computer-readable storage medium having a computer program stored thereon, the program implementing the method when executed by a processor.
[0016] By employing the above-described scheme, and through a scenario-based collaborative adjustment method and system for home environmental parameters, the following technical effects are achieved: 1) Enhanced privacy and security: Employs millimeter-wave radar to acquire human respiratory rhythm information without collecting image data, thus improving privacy and security.
[0017] 2) Precise scene determination: Through multimodal fusion, the system can accurately determine home scenes (such as sleep, exercise, and office) and reduce false triggers.
[0018] 3) Decoupling and coordination of multiple devices: Introduce an environmental parameter coordination and coupling model to resolve conflicts between multiple devices (such as the mutual exclusion between cooling and humidification) and maximize comfort.
[0019] 4) Dual closed-loop evolution: It not only relies on manual adjustment feedback, but can also perform imperceptible automatic correction based on the user's physiological indicators (such as chills, sweating, and shortness of breath), greatly reducing the frequency of manual operation.
[0020] 5) Edge computing data protection: Adopting a localized edge computing architecture, millimeter-wave radar point cloud data and model calculations are completed on the local gateway, without the need to upload raw data to the cloud, thus providing dual protection for user data privacy from both the physical and network layers.
[0021] The above description is merely an overview of the technical solution of the present invention. In order to better understand the technical means of the present invention and to implement it in accordance with the contents of the specification, the preferred embodiments of the present invention are described in detail below with reference to the accompanying drawings. Attached Figure Description
[0022] Figure 1It is a flowchart of the method for scene-based collaborative adjustment of home environment parameters according to the present invention; Figure 2 It is a logical structure diagram of the method for scene-based collaborative adjustment of home environment parameters in an embodiment of the present invention; Figure 3 It is a schematic diagram of the control timing in the sleep scene in the first embodiment of the present invention; Figure 4 It is a flowchart of the dual closed-loop adaptive adjustment based on user physiological feedback and manual intervention in an embodiment of the present invention. Detailed implementation manners
[0023] The following combines the drawings and embodiments to further describe the detailed implementation manners of the present invention. The following embodiments are used to illustrate the present invention but not to limit the scope of the present invention.
[0024] Refer Figure 1 As shown, this embodiment provides a method for scene-based collaborative adjustment of home environment parameters, including the following steps: Step S1: Collect the micro-motion signals in the home space through at least one millimeter-wave radar, perform phase demodulation and timing filtering on the micro-motion signals, and extract the human respiratory rhythm feature sequence; Step S2: Synchronously collect at least one environmental parameter in the home environment, where the environmental parameters include one or more of temperature, humidity, air quality parameters, and light intensity, and construct an environmental state vector; Step S3: Perform multi-modal fusion on the respiratory rhythm feature sequence and the environmental state vector, input it into the scene intention determination model, and output the scene label corresponding to the current home environment; Step S4: According to the scene label, call the environmental parameter collaborative coupling model to generate a multi-device control action vector, where the collaborative coupling model is configured to calculate the operating parameters of each environmental adjustment device based on preset device interaction constraint rules or multi-objective optimization functions; Step S5: Collaboratively control at least two environmental adjustment devices according to the multi-device control action vector; Step S6: Collect the manual intervention behaviors of the user on the environmental adjustment devices and the adjusted human physiological feedback indicators, and update the parameters of the environmental parameter collaborative coupling model based on the manual intervention behaviors or physiological feedback indicators to form a closed-loop optimization adjustment.
[0025] In this embodiment, the respiratory rhythm feature sequence at least includes respiratory frequency, respiratory amplitude stability, and micro-motion phase continuity indicators; the manual intervention behaviors include the opening, closing, or parameter adjustment operations of any environmental adjustment device in the automatic adjustment state; the physiological feedback indicators include the respiratory frequency change rate, body micro-vibration signal, and body movement amplitude.
[0026] In this embodiment, the millimeter-wave radar operates in a frequency band between 24 GHz and 77 GHz.
[0027] In this embodiment, the multimodal fusion in step S3 includes inputting the scene intent determination model after feature splicing or weighted fusion of the respiratory rhythm feature sequence and the environmental state vector.
[0028] In this embodiment, the scene intent determination model in step S3 is a supervised machine learning model or a deep learning model.
[0029] In this embodiment, the construction method of the environmental parameter cooperative coupling model in step S4 includes any of the following: (1) Linear weighted mapping: represented as Where A is the device control action vector, S is the scene state vector, and M is the environmental parameter cooperative coupling matrix. The element weights in the matrix are used to characterize the gain or suppression relationship between different devices. (2) Nonlinear mapping: Using a fuzzy logic controller or an artificial neural network (ANN), the scene labels and environmental state vectors are mapped to device control parameters.
[0030] In this embodiment, when using an environmental parameter co-coupling matrix, the weights are incrementally updated based on the frequency, direction, or magnitude of user intervention.
[0031] In this embodiment, the environmental control equipment includes two or more of the following: air purification equipment, humidification equipment, fragrance release equipment, and lighting equipment.
[0032] This embodiment also provides a scenario-based collaborative adjustment system for home environment parameters, including: Millimeter-wave sensing modules are used to collect micro-motion signals of the human body within the home space; The environmental parameter acquisition module is used to collect temperature, humidity, air quality, and light parameters; The scene intent determination module is used to perform multimodal fusion of human respiratory rhythm characteristics and environmental state and output scene labels; The collaborative control decision module is used to generate multi-device control action vectors based on a collaborative coupling model of environmental parameters. The feedback learning module is used to update the environmental parameter co-coupling model based on user manual intervention behavior and human physiological feedback indicators; Each module is configured to execute the method described.
[0033] Example 1: Collaborative Regulation of Nighttime Sleep Scenes Based on Breathing Rhythm 1. Sensing and Feature Extraction In a bedroom setting, a millimeter-wave radar is installed within 1.2-2.5 meters of the bedside to collect micro-motion signals of the human body at a sampling frequency of 20 Hz. Through phase demodulation and filtering, respiratory frequency, respiratory amplitude stability, and phase continuity indicators are extracted to construct a respiratory rhythm feature sequence.
[0034] 2. Construction of Environment State Vector Simultaneously collect environmental parameters and construct the following environmental state vector: Where T is temperature, RH is relative humidity, PM2.5 and VOC are air quality parameters, and L is light intensity.
[0035] 3. Scene Intent Determination The respiratory rhythm feature sequence is concatenated with the environmental state vector and input into a trained multimodal classification model. When the respiratory rate is detected to be between 12 and 16 breaths per minute, the light intensity is below 10 Lux, and the air quality parameters are within a stable range, the current scene is determined to be a deep sleep scene.
[0036] 4. Cooperative Coupling Model and Control Strategy For deep sleep scenarios, a linear weighted mapping model is adopted: Where A is the device control action vector, S is the scene state vector, and M is the environmental parameter collaborative coupling matrix, which is used to describe the control weight relationship between air purification, humidification, fragrance and lighting devices.
[0037] 5. Closed-loop optimization and update When a user manually adjusts any device in automatic adjustment mode, the system records the intervention and incrementally updates the corresponding weights in the collaborative coupling matrix.
[0038] Example 2: Collaborative Adjustment of Home Fitness Scenarios Based on Physiological Load Monitoring 1. Sensing and Dynamic Feature Extraction In the living room scenario, millimeter-wave radar detected large-amplitude, high-frequency human limb movements within the field of view (FOV). The system automatically switched radar processing modes from micro-motion detection mode to Doppler motion tracking mode. It extracted the target's motion amplitude (based on Doppler energy spectral density) and the real-time respiratory rate (respiratory rate > 25 breaths / min) after the movement. Combined with data on the increase in indoor carbon dioxide concentration, the scene intent determination model output a scene label as follows: "High-intensity home workout mode" .
[0039] 2. Cooperative coupling conflict handling In this scenario, the user's body surface temperature rises and cooling is required. However, if the air conditioner is directly turned on with strong cold wind blowing, it is easy to cause the user to catch a cold. Therefore, the environmental parameter cooperative coupling model calls the "exercise - comfort" constraint rule: Air - conditioning equipment: Set to "cooling mode", the wind direction is adjusted to "blow away from people" (based on the human body orientation determined by radar positioning), and the target temperature is set to 24°C.
[0040] Fresh - air equipment: The weight is adjusted to the highest level, and the "strong ventilation mode" is turned on to quickly reduce the carbon dioxide accumulation caused by exercise.
[0041] Humidification equipment: Temporarily turn off or reduce the power to assist in sweating and heat dissipation.
[0042] 3. Automatic closed - loop based on physiological feedback After the system adjusts and operates for 10 minutes, it continuously monitors the user's physiological indicators: If the radar monitors that the user's breathing frequency gradually decreases during the exercise - rest period, it indicates that the environmental load is appropriate, and the current parameters are maintained.
[0043] If the radar monitors that the user shows a micro - tremor signal (suspected shivering) or a hugging - arm action after stopping exercise, the system determines that the environmental temperature is too low, and immediately automatically raises the air - conditioner temperature and turns off the high - wind gear of the fresh air to achieve a non - perceptible correction.
[0044] As shown in Figure 2, the overall technical solution logic structure of a method for scenario - based cooperative regulation of home environmental parameters is as follows: Perception layer: Composed of millimeter - wave radar and environmental sensors, responsible for collecting human body micro - motion / breathing signals and environmental physical parameters (temperature / humidity / light / air quality) respectively.
[0045] Decision layer: Includes a multi - modal feature fusion module, a scene intention determination model, and an environmental parameter cooperative coupling model. This layer is responsible for converting the original signal into a scene label (such as "sleep"), and calculating the optimal device control vector.
[0046] Execution layer: Receives the control vector and drives the air purification, humidification, aroma, and lighting equipment to work cooperatively.
[0047] Feedback layer: Through monitoring the environmental change results and the user's manual intervention / physiological feedback, online updates the weights of the cooperative coupling model to form an adaptive closed - loop.
[0048] As shown in Figure 3 , the control timing in the sleep scenario of Embodiment 1 is as follows: T0-T1 (Preparation Period): When the user enters the bedroom and makes significant physical movements, the system enters "active mode," the lights are turned on, and the air purifier runs at full power.
[0049] T1-T2 (Sleep Onset Period): The user lies down, body movement decreases, and the light sensor detects the action of turning off the lights. The millimeter-wave radar begins to lock onto the breathing signal.
[0050] T2-T3 (Deep Sleep Determination): When the radar detects that the respiratory rate is stable at 12-16 breaths / min for a certain period of time, the system determines that the patient has entered a "deep sleep scenario".
[0051] T3-T4 (Coordinated Adjustment): The system automatically executes a sleep coordination strategy – the humidifier turns on to maintain comfort, the fragrance releases sleep-inducing scents, the air purifier switches to silent mode, and all indicator lights turn off.
[0052] T4-T5 (Closed-loop maintenance): The system continuously monitors physiological indicators and maintains the environment in the optimal sleep curve.
[0053] As shown in Figure 4, the dual closed-loop adaptive adjustment process based on user physiological feedback and human intervention demonstrates how the system monitors user overt behavior (human adjustment) and latent state (physiological indicators) in parallel after executing initial environmental control, and dynamically updates the weight parameters of the collaborative coupling model accordingly.
[0054] The present invention also provides a computer-readable storage medium having a computer program stored thereon, the program implementing the method when executed by a processor.
[0055] The present invention has the following technical effects: 1) Enhanced privacy and security: Employs millimeter-wave radar to acquire human respiratory rhythm information without collecting image data, thus improving privacy and security.
[0056] 2) Precise scene determination: Through multimodal fusion, the system can accurately determine home scenes (such as sleep, exercise, and office) and reduce false triggers.
[0057] 3) Decoupling and coordination of multiple devices: Introduce an environmental parameter coordination and coupling model to resolve conflicts between multiple devices (such as the mutual exclusion between cooling and humidification) and maximize comfort.
[0058] 4) Dual closed-loop evolution: It not only relies on manual adjustment feedback, but can also perform imperceptible automatic correction based on the user's physiological indicators (such as chills, sweating, and shortness of breath), greatly reducing the frequency of manual operation.
[0059] 5) Edge computing data protection: Adopting a localized edge computing architecture, millimeter-wave radar point cloud data and model calculations are completed on the local gateway, without the need to upload raw data to the cloud, thus providing dual protection for user data privacy from both the physical and network layers.
[0060] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the technical principles of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.
Claims
1. A scenario-based collaborative adjustment method for home environmental parameters, characterized in that, Includes the following steps: Step 1: Collect micro-motion signals in the home space using at least one millimeter-wave radar, perform phase demodulation and time-series filtering on the micro-motion signals, and extract the human respiratory rhythm feature sequence. Step 2: Simultaneously collect at least one environmental parameter in the home environment, including one or more of temperature, humidity, air quality parameters and light intensity, and construct an environmental state vector; Step 3: Perform multimodal fusion of the respiratory rhythm feature sequence and the environmental state vector, input the result into the scene intent determination model, and output the scene label corresponding to the current home environment; Step 4: Based on the scene label, call the environmental parameter collaborative coupling model to generate multi-device control action vectors, wherein the collaborative coupling model is configured to calculate the operating parameters of each environmental adjustment device based on preset device interaction constraint rules or multi-objective optimization functions. Step 5: Based on the multi-device control action vector, perform coordinated control on at least two types of environmental control devices; Step 6: Collect user's manual intervention behavior on the environmental regulation equipment and the human physiological feedback indicators after regulation, and update the parameters of the environmental parameter collaborative coupling model based on the manual intervention behavior or physiological feedback indicators to form a closed-loop optimization regulation.
2. The method for scenario-based coordinated adjustment of home environment parameters according to claim 1, characterized in that, The respiratory rhythm characteristic sequence includes at least respiratory rate, respiratory amplitude stability, and micro-motion phase continuity indicators; the artificial intervention behavior includes turning on, off, or adjusting the parameters of any environmental regulation device under automatic adjustment; the physiological feedback indicators include respiratory rate change rate, body micro-tremor signals, and body movement amplitude.
3. The method for scenario-based coordinated adjustment of home environment parameters according to claim 1, characterized in that, The millimeter-wave radar operates in the frequency band between 24 GHz and 77 GHz.
4. The method for scenario-based coordinated adjustment of home environment parameters according to claim 1, characterized in that, The multimodal fusion in step 3 includes inputting the scene intent determination model after feature concatenation or weighted fusion of the respiratory rhythm feature sequence and the environmental state vector.
5. The method for scenario-based coordinated adjustment of home environment parameters according to claim 4, characterized in that, The scene intent determination model mentioned in step 3 is a supervised machine learning model or a deep learning model.
6. The method for scenario-based coordinated adjustment of home environment parameters according to claim 1, characterized in that, The construction method of the environmental parameter cooperative coupling model in step 4 includes any of the following: (1) Linear weighted mapping: represented as Where A is the device control action vector, S is the scene state vector, and M is the environmental parameter cooperative coupling matrix. The element weights in the matrix are used to characterize the gain or suppression relationship between different devices. (2) Nonlinear mapping: Using a fuzzy logic controller or artificial neural network, the scene labels and environmental state vectors are mapped to the device control parameters.
7. The method for scenario-based coordinated adjustment of home environment parameters according to claim 6, characterized in that, When using an environmental parameter co-coupling matrix, the weights are incrementally updated based on the frequency, direction, or magnitude of user intervention.
8. The method for scenario-based coordinated adjustment of home environment parameters according to claim 1, characterized in that, The environmental control equipment includes two or more of the following: air purification equipment, humidification equipment, fragrance release equipment, and lighting equipment.
9. A scenario-based collaborative adjustment system for home environmental parameters, characterized in that, include: Millimeter-wave sensing modules are used to collect micro-motion signals of the human body within the home space; The environmental parameter acquisition module is used to collect temperature, humidity, air quality, and light parameters; The scene intent determination module is used to perform multimodal fusion of human respiratory rhythm characteristics and environmental state and output scene labels; The collaborative control decision module is used to generate multi-device control action vectors based on a collaborative coupling model of environmental parameters. The feedback learning module is used to update the environmental parameter co-coupling model based on user manual intervention behavior and human physiological feedback indicators; Each module is configured to perform the method described in any one of claims 1-8.
10. A computer-readable storage medium, characterized in that, It stores a computer program that, when executed by a processor, implements the method described in any one of claims 1-8.