A non-contact sensing-based virtual social emotion regulation rehabilitation cabin and method suitable for the elderly

Abstract

The application discloses a kind of based on non-contact sensing's old age adaptation virtual social emotion regulation rehabilitation cabin and method, wherein the rehabilitation cabin includes cabin body, non-contact physiological sensing module, is set to cabin body, for the physiological signal of user without contact acquisition;Virtual reality display module is set to cabin body inside top, presents three-dimensional virtual social scene;Audio interactive module;Communication module;Eye movement interactive module;Central control module.The application does not need old person to wear any equipment, realizes physiological and visual data without contact acquisition by the fusion perception of array millimeter wave radar sensor, miniature infrared surface array camera, effectively improves emotion recognition precision, greatly reduces use threshold, especially suitable for bedridden old people;According to fusion perception data real-time sensing emotion change, automatically adjust virtual environment parameter, while supporting old person voice / eye movement artificial interaction, realizes the closed-loop control of emotion intervention.

Description

Technical Field

[0001] This invention belongs to the field of virtual reality emotion regulation technology, specifically relating to an age-friendly virtual social emotion regulation and rehabilitation cabin and method based on non-contact perception. Background Technology

[0002] With an aging population, the psychological loneliness of bedridden elderly is becoming increasingly prominent. Due to limited mobility and reduced social interaction, elderly people who are bedridden for extended periods are prone to negative emotions such as loneliness, anxiety, and depression, which seriously affect their physical and mental health and quality of life.

[0003] There are some existing solutions that apply virtual reality technology to rehabilitation training. The VR rehabilitation equipment used in these technologies usually has the following shortcomings: (1) It requires users to wear sensors or operate controllers, which is a high threshold for the elderly to use; (2) The social scenes are mostly patient groups, lacking emotional connection with family members; (3) It cannot dynamically adjust the environment according to the elderly’s real-time psychological state, the interactive experience is mechanical and monotonous, and it is difficult to achieve personalized emotional regulation; (4) It lacks physiological abnormality warning and equipment failure emergency handling mechanisms, and its use is not safe enough; (5) It lacks age-friendly design details and does not consider the long-term physiological adaptation needs of bedridden elderly.

[0004] Based on the above reasons, an age-friendly virtual social emotion regulation and rehabilitation cabin and method based on non-contact perception are provided. Summary of the Invention

[0005] In view of the shortcomings of the existing technology, the purpose of this invention is to provide an age-friendly virtual social emotion regulation and rehabilitation cabin and method based on non-contact perception.

[0006] To achieve the above-mentioned technical objectives, the technical solution adopted by the present invention is as follows: In a first aspect, the present invention provides an age-friendly virtual social emotion regulation and rehabilitation cabin based on non-contact perception, comprising, The cabin consists of a main cabin, a forward cabin, and a rear cabin, both of which are retractable to the main cabin. The main cabin is equipped with an oxygen interface and a USB signal interface, and an emergency call button with a luminous indicator is installed inside the main cabin. The non-contact physiological sensing module is installed in the cabin and is used to collect the user's physiological signals without contact. The non-contact physiological sensing module adopts an array millimeter-wave radar sensor, a miniature infrared array camera and a signal preprocessing unit. The array millimeter-wave radar sensor is used to collect the user's heart rate variability, respiratory rate and body movement amplitude physiological signals without contact. The miniature infrared array camera is fixed on both sides of the top of the cabin and is used to assist in collecting the user's visual data, including facial micro-expression and limb micro-movement data. The signal preprocessing unit is used to filter out noise and extract physiological and visual feature parameters. The virtual reality display module is located on the top inside the cabin and presents a three-dimensional virtual social scene. The virtual reality display module includes an adjustable anti-glare VR display screen, and the audio interaction module includes directional speakers and array noise-canceling microphones. An audio interaction module for voice interaction, which includes directional speakers and an array of noise-canceling microphones; Eye-tracking interaction module; Communication module; The central control module is electrically connected to the non-contact physiological perception module, virtual reality display module, audio interaction module, communication module, and eye-tracking interaction module. The central control module includes an emotion recognition unit, a scene control unit, and a feedback optimization unit.

[0007] The cabin of this invention is placed on the bed where the elderly lie down. The elderly are the users. The side wall of the cabin is equipped with a transparent observation window, which makes it easy for caregivers to observe the elderly's condition inside the cabin. The cabin is made of medical-grade ABS+PC composite material, which has the characteristics of being breathable, scratch-resistant and easy to clean. Both the front and rear cabins can be stored behind the main cabin, reducing the space occupied when not in use and facilitating storage. This also allows bedridden elderly people to quickly leave the cabin without having to remove the entire cabin. The main cabin is equipped with an oxygen interface and a USB signal interface, and features an emergency call button with a luminous indicator. The oxygen interface supplies oxygen to the cabin, improving oxygen levels. The USB signal interface enables communication and control with the cabin. The emergency call button allows the elderly to call caregivers promptly, and the luminous indicator provides location guidance at night without creating excessive light that could disturb the elderly's sleep. The non-contact physiological sensing module uses an array millimeter-wave radar sensor, a miniature infrared array camera, and a signal preprocessing unit. The array millimeter-wave radar sensor is used to collect physiological signals such as heart rate variability, respiratory rate, and body movement amplitude of the elderly without contact. The miniature infrared array camera is fixed on both sides of the top of the cabin to assist in collecting the user's visual data. The visual data includes the elderly's facial micro-expression and limb micro-movement data. The two data are fused to improve the accuracy of emotion recognition and avoid the recognition error of a single sensor. The signal preprocessing unit is used to filter out noise and extract physiological and visual feature parameters. When the elderly are lying in bed, their line of sight is directed towards the virtual reality display module on the top of the cabin. The adjustable anti-glare VR display screen presents a three-dimensional virtual social scene. The display screen uses a double-layer anti-blue light coating and automatic brightness adjustment technology, which can automatically adjust the brightness according to the light intensity inside the cabin, filtering out more than 90% of harmful blue light and reducing eye fatigue for the elderly. The anti-glare VR display screen is installed by an electric bracket, which is used to adjust the angle of the anti-glare VR display screen to adapt to different lying positions. Through the eye-tracking interaction module, the elderly can issue various control commands by focusing and moving their eyes. The directional speaker in the audio interaction module uses ultrasonic array technology to create a sound field only inside the cabin, protecting the privacy of the elderly. The array noise-canceling microphone uses beamforming technology to filter out environmental noise and supports dialect recognition, ensuring accurate voice recognition for clear voice interaction. The communication module supports 5G and Wi-Fi 6, enabling low-latency connections with remote family members' terminals and allowing for synchronous interaction with virtual avatars. This allows seniors to engage in immersive family social interactions with their children and grandchildren in different locations. Additionally, the communication module supports the addition of large models, thereby enhancing the device's intelligence and functional diversity. The emotion recognition unit processes the data collected by the non-contact physiological sensing module and outputs emotion state labels, while judging whether the physiological signals are abnormal and triggering an early warning. The scene control unit retrieves the corresponding virtual social environment from the scene library based on the emotional state label and generates environment adjustment instructions; The feedback optimization unit builds personalized profiles for the elderly, records physiological data changes after each adjustment, and optimizes the parameters of the emotion recognition unit and the matching rules of the scene library.

[0008] In summary, the age-friendly virtual social emotion regulation and rehabilitation cabin of this invention requires no devices worn by the elderly. It achieves contactless collection of physiological and visual data through the fusion perception of array millimeter-wave radar sensors and miniature infrared array cameras, effectively improving the accuracy of emotion recognition and greatly reducing the barrier to use, making it particularly suitable for bedridden elderly. It can perceive emotional changes in real time based on fusion perception data and automatically adjust virtual environment parameters. At the same time, it supports voice / eye-tracking human interaction for the elderly, realizing closed-loop control of emotion intervention, with a high degree of intelligence. The cabin is designed for age-friendliness from multiple dimensions, including cabin material, physiological adaptation, visual protection (anti-blue light, automatic brightness), and ease of operation (electric support). It also has an emergency call button to improve the comfort and safety of the elderly. It builds personalized files for the elderly and continuously optimizes the emotion recognition model and scene library matching rules. With the increase in the number of uses, the emotion regulation effect continues to improve, realizing personalized services for each individual.

[0009] Secondly, this invention provides an age-friendly virtual social emotion regulation method based on non-contact perception, employing the aforementioned age-friendly virtual social emotion regulation rehabilitation cabin based on non-contact perception for virtual social emotion regulation, and including the following: S1. The non-contact physiological sensing module collects the user's physiological signal data and visual data in real time through the array millimeter-wave radar sensor and miniature infrared array camera, and filters out noise (bedding shaking, environmental vibration high-frequency noise) from the data. Specifically, the signal preprocessing unit in the non-contact physiological sensing module uses adaptive Kalman filtering, background modeling and foreground extraction methods to filter out noise in physiological and visual signals. At the same time, the fault self-testing unit monitors the working status of each module. If a fault is detected (such as a single radar sensor failure), a graded emergency handling is performed. If no fault is detected, the process continues to the next step. S2. The emotion recognition unit in the central control module extracts emotion feature parameters from the noise-filtered data and determines the current emotional state (loneliness index, anxiety index, calmness index). The emotion recognition unit uses a classification model based on deep neural networks to determine the user's current emotional state. The classification model based on deep neural networks is trained by pre-collected physiological signal samples of elderly people with different emotional states, different age groups, and different underlying diseases. Specifically, when the emotion recognition unit determines the current emotional state, it analyzes whether the collected physiological signals exceed the preset normal threshold. If they exceed the threshold, it triggers a physiological abnormality warning, sends an alarm to the caregiver's terminal, and suspends scene adjustment; if they do not exceed the threshold, it continues to the next step. S3. Based on emotional state, match an initial virtual social scene from a preset scene library, or dynamically adjust the current scene. The scene library includes at least one or more of the following virtual social scenes: virtual family gathering scene (the virtual family gathering scene supports remote family members to use lightweight motion capture on their smartphones to synchronize the expressions and simple movements of the virtual avatar), virtual familiar place roaming scene, and virtual leisure and entertainment scene. In addition, the dynamic adjustment of the current scene includes adjusting one or more of the following environmental parameters: the brightness of the virtual scene, color saturation, background music type, and the closeness of the virtual character; dynamic adjustment can also be triggered by the user's voice commands or eye control. S4. The adjusted virtual social environment is presented through the virtual reality display module, and the corresponding background sound and interactive voice are played through the audio interaction module; S5. Continuously monitor changes in the user's physiological signal data and visual data. When the user's emotional state is detected to improve to a preset threshold, maintain or enhance the current scene. The preset threshold is dynamically set based on the user's historical physiological data baseline. When the user's emotional state is detected to not improve or worsen, switch to another adjustment mode.

[0010] This invention designs virtual scenes with family social interaction at its core, allowing remote family members to synchronize virtual avatars via lightweight motion capture on smartphones. This enhances the elderly's sense of belonging, solves the problem of limited social scenarios in traditional devices, and addresses the technical issues of existing devices and methods being too difficult for bedridden elderly to use and unable to dynamically adjust the environment based on their real-time psychological state. It also adds physiological abnormality warnings and is compatible with nursing homes, rehabilitation centers, and home settings, enabling intelligent psychological care and family social support for bedridden elderly.

[0011] As a preferred technical solution of the present invention, in S1, the content of performing graded emergency handling includes: if a level 1 fault (such as a single radar sensor failure) is detected, the system switches to the backup perception mode to continue working and pushes a fault reminder; if a level 2 fault (such as a VR display screen or communication module failure) is detected, the system stops the virtual scene presentation and starts the basic monitoring mode.

[0012] This invention adds a fault self-checking unit to realize graded emergency handling, and sets up an early warning of abnormal physiological signals to promptly send alarms to nursing staff, thus solving the problem of insufficient safety in the use of existing equipment.

[0013] The beneficial effects of this invention are as follows: It eliminates the need for elderly individuals to wear any devices, achieving contactless collection of physiological and visual data through the fusion perception of an array of millimeter-wave radar sensors and a miniature infrared array camera. This effectively improves the accuracy of emotion recognition, significantly lowers the barrier to entry, and is particularly suitable for bedridden elderly individuals. Based on the fusion perception data, it perceives emotional changes in real time and automatically adjusts virtual environment parameters. It also supports voice / eye-tracking human-computer interaction (through the eye-tracking interaction module, the elderly can issue various control commands through eye gaze and line of sight movement), achieving closed-loop regulation of emotion intervention with a high degree of intelligence. The invention incorporates age-friendly design from multiple dimensions, including cabin material, physiological adaptation, visual protection (blue light protection, automatic brightness), and ease of operation (electric support). An emergency call button is also included to enhance the comfort and safety of elderly users. Furthermore, it constructs personalized profiles for each elderly individual, continuously optimizing the emotion recognition model and scene matching rules. With increased usage, the emotion regulation effect continuously improves, achieving personalized services tailored to each individual. Attached Figure Description

[0014] The present invention can be further illustrated by the non-limiting embodiments given in the accompanying drawings; Figure 1 This is a schematic diagram of the module composition of an embodiment of the present invention; Figure 2 This is a schematic diagram of an elderly person lying down while using the adjustable rehabilitation cabin in an embodiment of the present invention; Figure 3 This is a schematic diagram of an elderly person inside the cabin in an embodiment of the present invention; The symbols for the main components are explained below: 1. Cabin; 2. Non-contact physiological sensing module; 3. Virtual reality display module; 4. Audio interaction module; 5. Communication module; 6. Central control module; 7. Eye-tracking interaction module. Detailed Implementation

[0015] The technical solutions of the present invention will be described in detail below with reference to specific embodiments and accompanying drawings. The embodiments described herein are specific implementations of the present invention, used to illustrate the concept of the present invention; these descriptions are explanatory and exemplary, and should not be construed as limiting the implementation methods or the scope of protection of the present invention. In addition to the embodiments described herein, those skilled in the art can employ other obvious technical solutions based on the content disclosed in the claims and specification of this application. These technical solutions include those that make any obvious substitutions and modifications to the embodiments described herein. Example

[0016] like Figure 1 , 2 As shown in Figure 3, this embodiment provides an age-friendly virtual social emotion regulation and rehabilitation cabin based on non-contact perception, including: Cabin 1 includes a main cabin 11, a front cabin 12, and a rear cabin 13. Both the front cabin 12 and the rear cabin 13 are retractably connected to the main cabin 11. The main cabin 11 is equipped with an oxygen interface and a USB signal interface. An emergency call button with a luminous indicator is installed inside the main cabin 11. The non-contact physiological sensing module 2, located in the cabin 1, is used to collect the user's physiological signals without contact. The non-contact physiological sensing module 2 uses an array millimeter-wave radar sensor (three 60GHz millimeter-wave radar sensors, respectively aimed at the elderly person's chest, head, and limbs, with a detection accuracy of ±2bpm), a miniature infrared array camera (two in total, fixed on both sides of the top, with no visible light emission), and a signal preprocessing unit. The array millimeter-wave radar sensor is used to collect the user's heart rate variability, respiratory rate, and body movement amplitude physiological signals without contact. The miniature infrared array camera is fixed on both sides of the top of the cabin 1 to assist in collecting the user's visual data, including facial micro-expressions and limb micro-movement data. The signal preprocessing unit is used to filter out noise and extract physiological and visual feature parameters. The virtual reality display module 3 is located on the top inside the cabin 1 and presents a three-dimensional virtual social scene. The virtual reality display module 3 includes an adjustable anti-glare VR display screen, and the audio interaction module 4 includes a directional speaker and an array of noise-canceling microphones. Audio interaction module 4 is used for voice interaction. Audio interaction module 4 includes a directional speaker and an array of noise-canceling microphones. Communication module 5; Eye-tracking interaction module 7; The central control module 6 is electrically connected to the non-contact physiological perception module 2, the virtual reality display module 3, the audio interaction module 4, the communication module 5, and the eye-tracking interaction module 7. The central control module 6 includes an emotion recognition unit, a scene control unit, and a feedback optimization unit.

[0017] The cabin 1 of the present invention is placed on the bed where the elderly lie down. The elderly are the users. The side wall of the cabin 1 is provided with a transparent observation window, which makes it easy for caregivers to observe the elderly's condition inside the cabin. The cabin 1 is made of medical-grade ABS+PC composite material, which has the characteristics of breathability, scratch resistance and easy cleaning. Both the front cabin 12 and the rear cabin 13 can be stored behind the main cabin 11, reducing the space occupied when not in use and facilitating storage. This also allows bedridden elderly people to quickly leave the interior of the cabin 1 without having to remove the entire cabin 1. The main cabin 11 is equipped with an oxygen interface and a USB signal interface. It also has an emergency call button with a luminous indicator. The oxygen interface supplies oxygen to the interior of the cabin 1 to improve oxygen levels. The USB signal interface is used to communicate and control the cabin 1. The emergency call button allows the elderly to call caregivers in a timely manner. The luminous indicator provides location information at night without creating excessive light that could disturb the elderly's sleep. The non-contact physiological sensing module 2 uses an array millimeter-wave radar sensor, a miniature infrared array camera, and a signal preprocessing unit. The array millimeter-wave radar sensor is used to collect physiological signals such as heart rate variability, respiratory rate, and body movement amplitude of the elderly without contact. The miniature infrared array camera is fixed on both sides of the top of the cabin 1 to assist in collecting the user's visual data. The visual data includes the elderly's facial micro-expression and limb micro-movement data. The two data are fused to improve the accuracy of emotion recognition and avoid the recognition error of a single sensor. The signal preprocessing unit is used to filter out noise and extract physiological and visual feature parameters. When the elderly are lying in bed, their line of sight is directed towards the virtual reality display module 3 on the top of the inner side of the cabin 1. The adjustable anti-glare VR display screen presents a three-dimensional virtual social scene. The display screen uses a double-layer anti-blue light coating and automatic brightness adjustment technology, which can automatically adjust the brightness according to the light intensity inside the cabin, filter out more than 90% of harmful blue light, reduce eye fatigue for the elderly, and the anti-glare VR display screen is installed by an electric bracket, which is used to adjust the angle of the anti-glare VR display screen to adapt to different lying positions. The directional speaker in audio interaction module 4 uses ultrasonic array technology to make the sound field only inside the cabin, protecting the privacy of the elderly. The array noise-canceling microphone uses beamforming technology to filter out environmental noise and supports dialect recognition to ensure the accuracy of voice recognition, so as to achieve clear voice interaction. The communication module 5 supports 5G and Wi-Fi 6, which is used to establish low-latency connections with remote family members' terminals, enabling synchronous interaction with virtual avatars, allowing the elderly to have immersive family social interactions with their children and grandchildren in different places. The emotion recognition unit processes the data collected by the non-contact physiological sensing module 2 and outputs emotion state labels, while judging whether the physiological signals are abnormal and triggering an early warning. The scene control unit retrieves the corresponding virtual social environment from the scene library based on the emotional state label and generates environment adjustment instructions; The feedback optimization unit builds personalized profiles for the elderly, records physiological data changes after each adjustment, and optimizes the parameters of the emotion recognition unit and the matching rules of the scene library.

[0018] In summary, the age-friendly virtual social emotion regulation and rehabilitation cabin of this invention requires no devices worn by the elderly. It achieves contactless collection of physiological and visual data through the fusion perception of array millimeter-wave radar sensors and miniature infrared array cameras, effectively improving the accuracy of emotion recognition and greatly reducing the barrier to use, making it particularly suitable for bedridden elderly. It can perceive emotional changes in real time based on fusion perception data and automatically adjust virtual environment parameters. At the same time, it supports voice / eye-tracking human interaction for the elderly (using eye-tracking human interaction module 7) to achieve closed-loop control of emotion intervention, with a high degree of intelligence. The cabin is designed to be age-friendly from multiple dimensions, including cabin material, physiological adaptation, visual protection (anti-blue light, automatic brightness), and ease of operation (electric support). It also has an emergency call button to improve the comfort and safety of the elderly. It builds personalized files for the elderly and continuously optimizes the emotion recognition model and scene library matching rules. With the increase in the number of uses, the emotion regulation effect continues to improve, realizing personalized services for each individual. Example

[0019] This invention provides an age-friendly virtual social emotion regulation method based on non-contact perception, comprising the following contents: S1. The non-contact physiological sensing module 2 uses an array millimeter-wave radar sensor and a miniature infrared array camera to collect the user's physiological signal data and visual data in real time and filters out noise (bedding shaking, environmental vibration high-frequency noise). Specifically, the signal preprocessing unit in the non-contact physiological sensing module 2 uses adaptive Kalman filtering, background modeling and foreground extraction methods to filter out noise in physiological and visual signals. At the same time, the fault self-testing unit monitors the working status of each module. If a fault is detected (such as a single radar sensor failure), a graded emergency handling is performed. If no fault is detected, the process continues. S2. The emotion recognition unit in the central control module 6 extracts emotion feature parameters from the noise-filtered data and determines the current emotional state (loneliness index, anxiety index, calmness index). The emotion recognition unit uses a classification model based on deep neural networks to determine the user's current emotional state. The classification model based on deep neural networks is trained by pre-collected physiological signal samples of elderly people with different emotional states, different age groups, and different underlying diseases. Specifically, when the emotion recognition unit determines the current emotional state, it analyzes whether the collected physiological signals exceed the preset normal threshold. If they exceed the threshold, it triggers a physiological abnormality warning, sends an alarm to the caregiver's terminal, and suspends scene adjustment; if they do not exceed the threshold, it continues to the next step. S3. Based on emotional state, match an initial virtual social scene from a preset scene library, or dynamically adjust the current scene. The scene library includes at least one or more of the following virtual social scenes: virtual family gathering scene (the virtual family gathering scene supports remote family members to use lightweight motion capture on their smartphones to synchronize the expressions and simple movements of the virtual avatar), virtual familiar place roaming scene, and virtual leisure and entertainment scene. In addition, the dynamic adjustment of the current scene includes adjusting one or more of the following environmental parameters: the brightness of the virtual scene, color saturation, background music type, and the closeness of the virtual character; dynamic adjustment can also be triggered by the user's voice commands or eye control. In the virtual family gathering scenario, simulating a traditional family dinner during a festival, the elderly can see virtual avatars of their children and grandchildren sitting around the table; remote family members can use the front-facing camera and motion sensors of their smartphones to capture facial expressions, head movements, and simple hand gestures, which are then transmitted to the rehabilitation cabin with low latency via 5G / Wi-Fi 6, enabling synchronous interaction with the avatars; if family members do not have the means to capture motion, the virtual avatars can automatically match the facial expressions and body movements corresponding to the voice. In the virtual hometown tour scenario, three-dimensional reconstruction is carried out based on the places where the elderly lived when they were young (such as the streets of their hometown, the factories where they worked, etc.). The elderly can realize the virtual slow tour through eye control (optical sensors collect the trajectory of eye movement), or they can use voice commands to "move forward", "stay" and "switch perspective" to evoke fond memories. The virtual leisure and entertainment scene includes virtual teahouses, virtual parks, virtual theaters, etc. The elderly can have simple entertainment interactions with other virtual characters, such as playing chess, listening to opera, and chatting. The number and closeness of the virtual characters in the scene can be automatically adjusted according to the elderly’s emotional state. S4. The adjusted virtual social environment is presented through the virtual reality display module 3, and the corresponding background sound and interactive voice are played through the audio interaction module 4; S5. Continuously monitor changes in the user's physiological signal data and visual data. When the user's emotional state is detected to improve to a preset threshold, maintain or enhance the current scene. The preset threshold is dynamically set based on the user's historical physiological data baseline. When the user's emotional state is detected to not improve or worsen, switch to another adjustment mode.

[0020] In this embodiment, a virtual scene is designed with family social interaction as the core, supporting remote family members to achieve virtual avatar synchronization through lightweight motion capture on smartphones. This enhances the elderly's sense of belonging, solves the problem of limited social scenarios in traditional devices, and addresses the technical issues of high barriers to use for bedridden elderly and the inability to dynamically adjust the environment according to their real-time psychological state. It also adds a physiological abnormality warning function, adapting to nursing homes, rehabilitation centers, and home scenarios, enabling intelligent psychological care and family social support for bedridden elderly. Example

[0021] Training of emotion recognition unit model A hybrid architecture combining Convolutional Neural Networks (CNN) and Long Short-Term Memory (LSTM) is employed to significantly enhance generalization capabilities for older adults. 1. Training data collection: Recruit 50 elderly volunteers aged 60-85, of whom 40% have underlying diseases such as hypertension and diabetes. They watch different emotion-inducing videos in a laboratory environment. At the same time, physiological and visual data are collected through array millimeter-wave radar sensors and miniature infrared array cameras. Subjective emotion scores are obtained through questionnaires. The collected emotional states include four states: calm, joy, loneliness, and anxiety. 2. Data Preprocessing: Adaptive Kalman filtering and time-series smoothing are performed on the raw radar echo signal to extract feature parameters such as heart rate variability (HRV), respiratory rate (RR), and body energy (MA) (HRV includes SDNN, RMSSD time-domain indices and LF / HF frequency-domain indices); background modeling and foreground extraction are performed on the optical data to extract facial micro-expression and limb micro-movement features, and finally the two types of features are spatiotemporally aligned and fused. 3. Model Training: The preprocessed fused feature sequence was input into the CNN-LSTM network. The CNN layer extracted local temporal features, and the LSTM layer captured long-term dependencies. The model output was the probability distribution of four emotional states. The model was trained using the cross-entropy loss function and the Adam optimizer. The training set, validation set, and test set were divided in a 7:1:2 ratio. The overall test set accuracy was 89.3%, and the recognition accuracy for elderly people with different underlying diseases was no less than 85%. 4. Warning threshold setting: Based on training data, set warning thresholds for physiological abnormalities. A warning will be triggered when the heart rate is >120 bpm or <50 bpm, or the respiratory rate is >30 breaths / minute or <10 breaths / minute. Example

[0022] Taking an 85-year-old bedridden patient with hypertension using the emotion regulation rehabilitation cabin of this invention as an example, the complete workflow is explained as follows: 1. The rehabilitation cabin is set up on the bed where the elderly person is lying. When the system is started, the non-contact physiological sensing module 2 begins to collect physiological and visual data. The fault self-testing unit completes the status detection of each module. If there is no fault, it enters the normal working mode. 2. In the initial stage, the elderly person's heart rate was 78 bpm, respiratory rate was 20 breaths / minute, body movement frequency was high, and facial micro-expression was frowning. After the emotion recognition unit fused the data, it judged that the elderly person was in an anxious state. The physiological signals did not exceed the warning threshold, so the elderly person entered the scene adjustment stage. 3. Based on the anxiety state, the scene control unit selects the "tranquil lakeside" virtual scene from the scene library. The virtual reality display module presents a picture of the shimmering lake surface, the audio interaction module 4 plays soft sounds of flowing water and birdsong, the light brightness is adjusted to 80 nits, and the color saturation is reduced to 60%. 4. After the elderly person sees the scene through the VR display screen, their breathing gradually stabilizes within 5 minutes, their heart rate drops to 72 bpm, their face relaxes, and the system detects that their emotional state has improved to the calm threshold, and automatically switches the scene to the "family living room" virtual scene.

[0023] 5. Communication module 5 establishes a connection with the daughter's mobile app in a distant location. The daughter uses her mobile phone to capture facial expressions and head movements, and her virtual avatar appears in the living room screen. The elderly man conducts real-time voice calls with his daughter through a noise-canceling microphone (supporting dialects). During the call, the elderly man's heart rate is stable at around 70 bpm, and his emotional state is calm. 6. After the call ends, the elderly person exits the scene via voice command. The system stops the virtual scene presentation and at the same time, the feedback optimization unit records the data of the entire interaction process (initial anxiety state, lakeside scene adjustment parameters, emotion improvement curve, positive response in the family living room scene), increases the matching priority of the "family living room" scene, and fine-tunes the feature weights of the elderly person's "frowning" micro-expression and anxiety state in the emotion recognition model. 7. The system will revert to basic monitoring mode and continue to collect physiological signals from the elderly. If no abnormalities are detected, it will automatically enter low-power mode after 30 minutes.

[0024] The above embodiments are merely illustrative of the principles and effects of the present invention and are not intended to limit the present invention. Any person skilled in the art can modify or alter the above embodiments without departing from the spirit and scope of the present invention. Therefore, all equivalent modifications or alterations made by those skilled in the art without departing from the spirit and technical concept disclosed in the present invention should still be covered by the claims of the present invention.

Claims

1. An age-friendly virtual social emotion regulation and rehabilitation cabin based on non-contact perception, characterized in that: include, hull (1); A non-contact physiological sensing module (2) is installed in the cabin (1) for collecting the user's physiological signals without contact. The virtual reality display module (3) is located on the top of the inner side of the cabin (1) and presents a three-dimensional virtual social scene; Audio interaction module (4), used for voice interaction; Communication module (5); Eye-tracking interaction module (7); The central control module (6) is electrically connected to the non-contact physiological perception module (2), the virtual reality display module (3), the audio interaction module (4), the communication module (5), and the eye-tracking interaction module (7).

2. The age-friendly virtual social emotion regulation and rehabilitation cabin based on non-contact perception according to claim 1, characterized in that: The non-contact physiological sensing module (2) uses an array millimeter-wave radar sensor, a miniature infrared array camera and a signal preprocessing unit. The array millimeter-wave radar sensor is used to collect physiological signals such as heart rate variability, respiratory rate and body movement amplitude of the user without contact. The miniature infrared array camera is fixed on both sides of the top of the cabin (1) to assist in collecting the user's visual data, including facial micro-expression and limb micro-movement data. The signal preprocessing unit is used to filter out noise and extract physiological and visual feature parameters. The virtual reality display module (3) includes an adjustable-angle anti-glare VR display, and the audio interaction module (4) includes a directional speaker and an array of noise-canceling microphones.

3. A virtual social emotion regulation and rehabilitation cabin for the elderly based on non-contact perception according to claim 1, characterized in that: The cabin (1) includes a main cabin (11), a front cabin (12), and a rear cabin (13). Both the front cabin (12) and the rear cabin (13) are retractable and connected to the main cabin (11).

4. A virtual social emotion regulation and rehabilitation cabin for the elderly based on non-contact perception according to claim 3, characterized in that: The main cabin (11) is equipped with an oxygen interface and a USB signal interface. An emergency call button with a night light indicator is installed inside the main cabin (11).

5. A virtual social emotion regulation and rehabilitation cabin for the elderly based on non-contact perception according to claim 1, characterized in that: The central control module (6) includes an emotion recognition unit, a scene control unit, and a feedback optimization unit; The emotion recognition unit processes the data collected by the non-contact physiological sensing module (2) and outputs the emotion status label, while judging whether the physiological signal is abnormal and triggering an early warning. The scene control unit retrieves the corresponding virtual social environment from the scene library based on the emotional state label and generates environment adjustment instructions; The feedback optimization unit builds personalized profiles for the elderly, records physiological data changes after each adjustment, and optimizes the parameters of the emotion recognition unit and the matching rules of the scene library.

6. A method for age-friendly virtual social emotion regulation based on non-contact perception, characterized in that: The invention employs a non-contact sensing-based, age-friendly virtual social emotion regulation and rehabilitation cabin according to any one of claims 1 to 5 for virtual social emotion regulation, and includes the following contents: S1. The user's physiological signal data and visual data are collected in real time through the non-contact physiological sensing module (2), and the noise of the data is filtered out. S2. The central control module (6) extracts the emotional feature parameters from the noise-filtered data and determines the current emotional state; S3. Based on emotional state, match an initial virtual social scene from a preset scene library, or dynamically adjust the current scene; S4. The adjusted virtual social environment is presented through the virtual reality display module (3), and the corresponding background sound and interactive voice are played through the audio interaction module (4); S5. Continuously monitor changes in the user's physiological signal data and visual data. When the user's emotional state is detected to improve to a preset threshold, maintain or enhance the current scene. The preset threshold is dynamically set based on the user's historical physiological data baseline. When the user's emotional state is detected to not improve or worsen, switch to another adjustment mode.

7. A method for age-friendly virtual social emotion regulation based on non-contact perception according to claim 6, characterized in that: In S2, the central control module (6) determines whether the collected physiological signals exceed the preset normal threshold. If they do, it triggers a physiological abnormality warning, sends an alarm to the nursing staff terminal, and suspends scene adjustment. If they do not exceed the threshold, it proceeds to S3.

8. A method for age-friendly virtual social emotion regulation based on non-contact perception according to claim 6, characterized in that: In S2, the central control module (6) uses a classification model based on deep neural networks to determine the user's current emotional state. The classification model based on deep neural networks was trained using physiological signal samples of elderly people with different emotional states, age groups, and underlying diseases collected in advance.

9. A method for age-friendly virtual social emotion regulation based on non-contact perception according to claim 6, characterized in that: In S3, the scene library includes at least one or more of the following virtual social scenes: virtual family gathering scene, virtual home tour scene, and virtual leisure and entertainment scene.

10. A method for age-friendly virtual social emotion regulation based on non-contact perception according to claim 6, characterized in that: In S3, dynamic adjustment of the current scene includes adjusting one or more of the following environmental parameters: the brightness of the virtual scene, color saturation, background music type, and the proximity of the virtual character; dynamic adjustment can also be triggered by the user's voice commands or eye control.

Images