Interactive toy for cultivating psychological toughness of teenagers

Abstract

The invention discloses an interactive toy for training psychological toughness of teenagers in the technical field of psychological training. The interactive toy comprises a VR glasses body used for providing a virtual reality training scene for a user; the first acquisition unit is used for acquiring facial muscle pressure change data of an area, shielded by the VR glasses, of a user; the second acquisition unit is used for synchronously acquiring facial expression dynamic data of an unshielded area of the user; the processing unit is used for establishing a personalized psychological state model of the user; acquiring dynamic data of the first acquisition unit and the second acquisition unit in real time, inputting the dynamic data into the personalized psychological state model, evaluating the current psychological state of the user, and generating an adjustment instruction according to an evaluation result; and the adjusting unit is used for adjusting the content presented in the VR glasses body in real time based on the adjusting instruction so as to match the current psychological state of the user. According to the scheme, the adverse effect on the training result caused by the fact that the facial expression cannot be completely captured due to the fact that the VR glasses shield the face can be reduced, and the training effect is improved.

Description

Technical Field

[0001] The invention belongs to the technical field of psychological training, and in particular is an interactive toy for cultivating the psychological resilience of teenagers. Background Art

[0002] Adolescence is a critical period of physical and psychological development, during which individuals face numerous complex factors, including academic pressures, interpersonal challenges, and family expectations. Psychological resilience is the ability to adapt and recover effectively under pressure from adversity, setbacks, and trauma. Adolescents with high levels of psychological resilience are more proactive in coping with life's challenges, maintaining good mental health and reducing the risk of psychological problems such as anxiety and depression. This lays the foundation for their social adaptation and personal growth in adulthood.

[0003] Traditional approaches to cultivating resilience, such as psychological counseling, educational seminars, and courses, have limitations. While counseling can provide professional assistance, access to services is limited, and some young people may be reluctant to seek help due to misunderstandings or resistance. Educational seminars and courses often involve a one-way flow of knowledge, which can be difficult to capture the deep interest and resonance of young people, leading to difficulties in applying what they've learned to cope with stress and setbacks in real life.

[0004] There are already some interactive toys or equipment on the market that use VR technology for adolescent psychological training. These products combine immersive experience with scientific intervention methods (such as cognitive behavioral therapy) to help adolescents manage emotions, train social skills, and release stress. For example, a VR mental health all-in-one machine integrates high-tech technologies such as virtual reality technology, three-dimensional panorama, and three-dimensional modeling. It can adjust the difficulty according to user behavior feedback, such as gradually increasing challenges in anxiety treatment scenarios. However, these devices have a significant flaw: trainees need to be equipped with VR glasses for training, and VR glasses block part of the face, making it impossible to accurately capture complete expressions, thereby affecting the accuracy of feedback and, in turn, affecting the training effect. If the VR glasses are removed to capture the complete expression, the immersive experience will be destroyed, which is also not conducive to improving the training effect.

[0005] In view of the defects of existing technologies, there is an urgent need for a new interactive toy that can not only give full play to the immersive experience advantages of VR glasses, but also effectively solve the contradiction that expressions cannot be fully captured due to VR glasses blocking the face, thereby improving the effect of training for adolescent psychological resilience. Summary of the Invention

[0006] In order to solve the above problems, the purpose of the present invention is to provide an interactive toy for cultivating psychological resilience in adolescents, which can effectively reduce the adverse effects of the inability to fully capture expressions due to VR glasses blocking the face on the training results, thereby improving the quality and effect of psychological resilience training.

[0007] In order to achieve the above object, the technical solution of the present invention is as follows:

[0008] An interactive toy for cultivating psychological resilience in adolescents, comprising a VR glasses body, a first acquisition unit, a second acquisition unit, a processing unit, and an adjustment unit;

[0009] The VR glasses are used to provide users with virtual reality training scenes;

[0010] The first acquisition unit is integrated into the VR glasses body and is used to collect the facial muscle pressure change data of the user in the area blocked by the VR glasses;

[0011] The second acquisition unit is used to synchronously acquire dynamic facial expression data of the user's unobstructed area;

[0012] The processing unit is configured to establish a personalized psychological state model of the user based on the correlation between the muscle pressure data of the occluded area and the facial expression data of the unoccluded area; and when the user is performing VR training, acquire dynamic data from the first acquisition unit and the second acquisition unit in real time, input the dynamic data into the personalized psychological state model, evaluate the user's current psychological state, and generate adjustment instructions based on the evaluation results;

[0013] The adjustment unit is used to adjust the content presented in the VR glasses body in real time based on the adjustment instructions to match the user's current psychological state.

[0014] Principle: When a teenager uses VR glasses for virtual training, the first and second acquisition units respectively capture data on the user's various facial expressions in response to various voice commands, building a personalized psychological state model of the user. During normal training, the data captured in real time by the first and second acquisition units is input into the personalized psychological state model. The processing unit determines the user's emotional changes in the current virtual scene content and assesses whether the teenager is currently in an anxious, focused, relaxed, or other psychological state. Based on the processing unit's instructions, the adjustment unit adjusts the difficulty of the virtual scene presented in the VR glasses, the interactive plot, and the physical interaction methods of the toy body in real time. For example, if the teenager is detected to be anxious due to the difficulty of the scene, and the changes in expression and facial muscle pressure indicate high stress, the adjustment unit can reduce the complexity of the task in the virtual scene or change the task type, introducing some relatively simple auxiliary training steps to guide the teenager to gradually build confidence. If the teenager is found to be relaxed and focused, and has the energy to undergo more intensive training, the scene difficulty and interaction complexity can be appropriately increased to further strengthen their psychological resilience, thereby achieving personalized and adaptive psychological resilience training results.

[0015] The above scheme has the following beneficial effects:

[0016] 1. This solution uses a first acquisition unit to capture facial muscle pressure change data in areas obscured by VR glasses, and a second acquisition unit to simultaneously capture dynamic facial expression data in unobstructed areas. This allows for comprehensive capture of adolescents' facial expressions during training. Based on this data, the processing unit builds a personalized psychological state model and assesses psychological states in real time, enabling the adjustment unit to accurately adjust the virtual scene to the user's psychological state. This creates a closed "assessment-adjustment-reassessment" training loop, meeting adolescents' needs for developing personalized psychological resilience.

[0017] 2. This solution can capture complete facial expression data without removing VR glasses, ensuring that young people can fully devote themselves to training in a highly immersive virtual environment while obtaining accurate feedback information in real time, maintaining the continuity and integrity of the training process, greatly improving training enthusiasm and participation, and ensuring training effectiveness.

[0018] 3. This program, through timely adjustments to training content and methods, helps young people better cope with the challenges and pressures of virtual scenarios, enabling them to more calmly face difficulties and setbacks in real life, thereby enhancing their psychological resilience and mental health. This program reduces the risk of anxiety, depression, and other psychological problems in young people, laying a solid foundation for their social adaptation and personal growth in adulthood, and positively promoting their mental health and social development.

[0019] Furthermore, the processing unit determines the user's current mental state through the following logic:

[0020] a. Data alignment and verification: comparing real-time facial muscle pressure change data with the preset pressure threshold range in the model. If the threshold is exceeded, an emotional abnormality flag is triggered;

[0021] b. Expression feature matching: extract key parameters from real-time facial expression dynamic data and perform similarity matching with the emotion feature library in the model;

[0022] c. Comprehensive assessment decision-making, combining the abnormal muscle pressure mark and the expression matching results. If both point to the same emotional type, it is determined to be that emotional state; if there is a conflict, the priority is determined according to the preset weight coefficient.

[0023] Beneficial effects: It can more accurately and comprehensively judge the user's psychological state, effectively avoid misjudgment caused by a single data source, and improve the pertinence and effectiveness of psychological resilience training.

[0024] Furthermore, the adjustment instructions are adapted to the user's psychological state through the following logic:

[0025] If it is determined to be an anxious state, the complexity of the scene task is reduced and guiding positive feedback is inserted;

[0026] If the player is judged to be in a focused state, the challenge difficulty will be gradually increased, and multi-threaded tasks will be added to enhance the player's ability to withstand stress;

[0027] If it is determined to be a relaxed state, the current scene parameters will be maintained and the user's immersive experience will be prolonged to consolidate the effect of cultivating psychological resilience.

[0028] Beneficial effects: Dynamically adjust VR training scenes according to the user's different psychological states to achieve personalized training, effectively relieve anxiety, improve concentration and stress resistance, consolidate the effect of psychological resilience training, improve the pertinence and effectiveness of training, and promote the development of psychological resilience in adolescents.

[0029] Furthermore, the steps of constructing the personalized mental state model include:

[0030] S1: Collect user baseline data through a preset standardized emotion induction experiment;

[0031] S2: Perform time-frequency domain analysis on the facial muscle pressure change data in the occluded area to extract the pressure distribution pattern and dynamic change gradient features; extract geometric features and dynamic features of the facial expression dynamic data in the unoccluded area through a lightweight convolutional network;

[0032] S3: Establish the mapping relationship between muscle pressure features and facial expression features, and generate the correlation weight matrix;

[0033] S4: Train the CNN-LSTM hybrid network through supervised learning and output a user-specific emotion classifier.

[0034] Beneficial Effects: Combining time-frequency domain analysis with a lightweight convolutional network, accurate facial features can be extracted from both occluded and unoccluded areas. This allows for a highly accurate emotion classifier to be trained using a CNN-LSTM hybrid network, enabling more accurate identification of user emotions. A personalized psychological state model is constructed based on the user's unique emotion classifier, enabling accurate assessment of the user's psychological state. This allows VR training scenarios to better align with the user's actual situation, improving the relevance and effectiveness of psychological resilience training.

[0035] Furthermore, it also includes a third acquisition unit and an optimization unit; the third acquisition unit is used to collect the user's heart rate and respiratory rate data; the optimization unit is used to jointly analyze the heart rate and respiratory rate data with the evaluation results of the processing unit based on the biofeedback mechanism, and dynamically optimize the scene adjustment strategy.

[0036] Beneficial effects: By introducing heart rate and respiratory rate data and combining it with the biofeedback mechanism, it can more comprehensively reflect the user's psychological and physiological state, thereby more accurately assessing the user's psychological state and avoiding misjudgment caused by a single data source.

[0037] The optimization unit dynamically adjusts scene adjustment strategies based on heart rate and respiratory rate data, making VR training scene adjustments more personalized and precise. For example, when a user is under great psychological pressure, even if their facial expressions do not clearly indicate anxiety, changes in heart rate and respiratory rate can trigger appropriate scene adjustments, reducing task difficulty or providing guiding feedback to help the user relieve stress.

[0038] Furthermore, the first acquisition unit includes a plurality of flexible piezoresistive sensor arrays, which are embedded in the inner contact surface of the VR glasses body in a matrix form.

[0039] Beneficial effects: The flexible piezoresistive sensor has good flexibility and fit. After being embedded in the inner contact surface of VR glasses, it will not cause discomfort to the user, ensuring that the user can maintain a comfortable wearing experience during long-term use, which is conducive to young people to better participate in psychological resilience training.

[0040] Furthermore, the first acquisition unit covers the user's cheekbone, glabella, nose bridge, frontalis muscle and temporalis muscle area.

[0041] Beneficial effects: The above-mentioned areas are key parts of facial expression changes. Covering these areas can more accurately capture the changes in facial muscle pressure of users in different emotional states, provide more representative data for psychological state assessment, and improve the accuracy of assessment.

[0042] Furthermore, the second acquisition unit is a camera, which is arranged on the side of the VR glasses body close to the user's unobstructed face.

[0043] Beneficial Effects: The camera can clearly capture the details of the user's facial expressions that are not obscured by VR glasses, providing high-quality dynamic expression data and ensuring the accuracy and reliability of expression feature extraction. Positioned close to the unobstructed face, the camera can precisely focus on the target area, ensuring that the collected data is synchronized in time and space with the occluded area data from the first acquisition unit, improving the overall accuracy of psychological state assessment.

[0044] Furthermore, it also includes a feedback unit; the feedback unit is used to output corresponding encouragement or guidance information to the user based on the user's performance in VR training and the psychological state evaluation results.

[0045] Beneficial effects: Through the personalized feedback mechanism, it can better meet the needs of different users in different psychological states and improve the effectiveness and pertinence of psychological resilience training.

[0046] Furthermore, it also includes a fourth acquisition unit and an abnormal data screening unit; the fourth acquisition unit is used to collect the vibration frequency of the user during VR training; the abnormal data screening unit is used to judge the user's current state based on the user's vibration frequency during VR training. If the vibration frequency exceeds a preset threshold, it is judged whether the user's current state is caused by emotional fluctuations. If so, the data collected by the first acquisition unit and the second acquisition unit are marked as valid data related to emotional fluctuations, otherwise, they are marked as interference data.

[0047] Beneficial effects: By monitoring vibration frequency and filtering out valid data caused by emotional fluctuations, misjudgments caused by non-emotional factors (such as accidental movements or equipment interference) are avoided, thereby improving the accuracy of the data based on which the psychological state assessment is based.

[0048] Furthermore, it also includes a fifth collection unit and a verification unit; the fifth collection unit is used to collect sound data emitted by the user; the verification unit is used to verify the accuracy of the interference data judgment based on the sound data.

[0049] Beneficial effects: By introducing sound data as a verification basis, it is possible to more accurately distinguish between valid data and interference data caused by emotional fluctuations, reduce misjudgments, and improve the accuracy of data screening.

[0050] Furthermore, the abnormal data screening unit analyzes and determines the user's current status based on the following logic:

[0051] Setting a preset threshold range for the vibration frequency, and triggering an abnormal state determination process when the vibration frequency collected by the fourth collection unit exceeds the preset threshold range;

[0052] After the abnormality judgment process is triggered, the vibration frequency data is first analyzed in the time domain and frequency domain to extract the characteristic parameters of the vibration frequency, including but not limited to the amplitude, frequency range, duration and frequency change rate of the vibration frequency;

[0053] Retrieving a pre-built first cough-related characteristic parameter database, which stores vibration frequency characteristic parameters corresponding to coughs of different types and intensities, the extracted vibration frequency characteristic parameters are compared with the vibration frequency characteristic parameters corresponding to coughs in the first characteristic parameter database, and the similarity between the two is calculated. If the similarity reaches or exceeds a preset cough matching threshold, the current vibration frequency phenomenon is determined to be caused by coughing, and the data currently collected by the first and second collection units is marked as interference data caused by coughing;

[0054] If the similarity with the cough characteristic parameter is lower than a preset cough matching threshold, a pre-established second characteristic parameter database related to emotional fluctuations is further retrieved. The second characteristic parameter database stores vibration frequency characteristic parameters corresponding to different emotional fluctuations. The extracted vibration frequency characteristic parameters are also compared with the vibration frequency characteristic parameters corresponding to emotional fluctuations in the second characteristic parameter database to calculate the similarity. If the similarity reaches or exceeds the preset emotional fluctuation matching threshold, it is determined that the current vibration frequency phenomenon is caused by emotional fluctuations, and the data currently collected by the first and second collection units are marked as valid data related to emotional fluctuations, which is used for subsequent comprehensive assessment of the user's psychological state in combination with other collected data.

[0055] If the similarity with the cough and emotional fluctuation characteristic parameters does not reach the corresponding matching threshold, the data currently collected by the first acquisition unit and the second acquisition unit will be temporarily stored as pending data, and the changes in the subsequently collected vibration frequency data will continue to be observed. When the subsequent continuously collected vibration frequency data are normal and the abnormal judgment process is not triggered again, the pending data will be regarded as accidental interference data and eliminated; if the abnormal judgment process is triggered again and the feature similarity with the pending data exceeds the preset similarity threshold, the two will be combined and the above comparison analysis will be performed again to determine the final cause.

[0056] Beneficial effects: Different from traditional single data judgment, this solution innovatively integrates vibration frequency data, facial muscle pressure change data and facial expression dynamic data, combines preset threshold range, feature parameter extraction, and multi-database comparison to form a unique comprehensive judgment logic, effectively screening out interference data caused by coughing, breaking through the limitation of existing technology that it is difficult to distinguish between coughing and normal emotional fluctuation data, and providing a more accurate data source for psychological state assessment.

[0057] Furthermore, the verification unit performs verification based on the following logic:

[0058] extracting characteristic parameters of the sound data collected by the fifth collection unit;

[0059] Comparing the characteristic parameters of the sound data with the characteristic parameters in a cough sound characteristic database constructed in advance;

[0060] If the similarity between the sound data and the cough sound characteristics reaches or exceeds the preset cough sound matching threshold, it is determined that the current vibration frequency phenomenon is caused by coughing, and the relevant data is confirmed to be interference data; otherwise, the interference data is changed to valid data.

[0061] Beneficial Effects: The dual verification mechanism significantly improves the accuracy of interference data screening, providing a purer and more reliable data source for psychological state assessment, effectively avoiding misjudgments of psychological state due to coughing. It also enhances the system's intelligent adaptation capabilities to different user scenarios, ensuring that the data used for psychological state assessment is always of high quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] Figure 1 This is a schematic diagram of the three-dimensional structure of the interactive toy for cultivating psychological resilience in teenagers according to the present invention.

[0063] Figure 2 This is a system block diagram of the interactive toy for cultivating psychological resilience in teenagers according to the present invention.

[0064] The reference numerals in the drawings of the specification include: 1. VR glasses body; 2. surrounding band; 3. headband; 4. first piezoresistive sensor; 5. second piezoresistive sensor. DETAILED DESCRIPTION

[0065] The following describes embodiments of the present invention in detail. Examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended only to explain the present invention and are not to be construed as limiting the present invention.

[0066] In the description of the present invention, it should be understood that the terms "longitudinal", "transverse", "vertical", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", etc., indicating the orientation or position relationship, are based on the orientation or position relationship shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present invention.

[0067] In the description of the present invention, unless otherwise specified and limited, it should be noted that the terms "installed", "connected" and "connected" should be understood in a broad sense. For example, it can be a mechanical connection or an electrical connection, or it can be the internal communication between two components. It can be a direct connection or an indirect connection through an intermediate medium. For ordinary technicians in this field, the specific meanings of the above terms can be understood according to the specific circumstances.

[0068] The following is further described in detail through specific implementation methods:

[0069] The embodiment is basically as shown in the attached Figure 1-Figure 2 Shown: An interactive toy for cultivating psychological resilience in adolescents, mainly including the following components:

[0070] VR glasses body 1, as the visual presentation core of the entire interactive toy, can create immersive virtual reality training scenes for users. These scenes are designed for psychological resilience training, such as simulating campus social situations, coping with exam pressure, dealing with frustration and other situations, to help teenagers develop psychological resilience in a virtual environment. The optical system of VR glasses body ensures clear images and a wide field of view, combined with the attached Figure 1 As shown, the VR glasses body 1 is equipped with a wearable component. In this embodiment, the wearable component includes a headband 3 and a circumferential strap 2. The headband 3 and the circumferential strap 2 are made of elastic material, which can adapt to the head shape and facial features of teenagers, ensuring that the user will not feel uncomfortable during long-term use.

[0071] The first acquisition unit, integrated into the VR glasses body 1, consists of a plurality of flexible piezoresistive sensor arrays, including a first piezoresistive sensor 4 and a second piezoresistive sensor 5. These sensors are precisely embedded in the inner contact surface of the VR glasses body in a matrix form and fully cover key areas such as the user's cheekbones, glabella, nose bridge, frontal muscles, and temporalis muscles. They can monitor and collect data on changes in facial muscle pressure in areas blocked by the VR glasses in real time, providing an important basis for subsequent analysis of the user's emotional state. For example, the first piezoresistive sensor 4 installed on the headband 3 is used to collect pressure changes in the user's frontal muscle area (the connection between pressure changes in areas such as the cheekbones, glabella, nose bridge, frontal muscles, and temporalis muscles and changes in facial expression is obtained through experimental testing). The second piezoresistive sensor 5 installed on the inside of the band 2 is used to collect pressure changes in the user's temporalis muscle area. The remaining piezoresistive sensors are arranged in the VR glasses body 1, specifically on the contact surfaces with the user's cheekbones, glabella, nose bridge, and other areas.

[0072] The second acquisition unit uses a high-definition camera and is installed on the side of the VR glasses body close to the user's unobstructed face. The camera has an autofocus function to ensure that facial expressions can always be clearly captured during the user's activities. Its high resolution and high frame rate characteristics make the captured facial expression dynamic data more delicate and smooth, and can accurately record every expression change of the user during the training process, such as the user's mouth corners, eye corners, forehead and other unobstructed areas. These dynamic data are compressed and stored in a specific encoding format (such as AU4 frowning, AU12 mouth corners raised) so that subsequent processing units can quickly read and analyze them. Both the first acquisition unit and the second acquisition unit use hardware clock synchronization technology (using the IEEE 1588 precise clock protocol) to ensure millisecond-level timing consistency between muscle pressure data and expression data.

[0073] The processing unit, as the "brain" of the entire interactive toy, uses a high-performance multi-core processor with powerful data processing and parallel computing capabilities. Its main functions include:

[0074] A personalized user psychological state model is established based on the correlation between muscle pressure data in the occluded area and facial expression data in the unoccluded area. Baseline user data is collected through a pre-set standardized emotion induction experiment. This data covers the user's typical facial muscle pressure and facial expression characteristics under different emotional states. (For example, when the user first wears the device, a calibration procedure (such as biting movements and standard smiles) is used to establish a mapping relationship between facial muscle pressure distribution and facial expression movements.) The processing unit performs time-frequency domain analysis on the facial muscle pressure change data in the occluded area, extracting pressure distribution patterns (such as the positive correlation between peak pressure in the zygomatic region and anxiety) and dynamic gradient characteristics (such as the instantaneous rate of increase in corrugator muscle pressure after task failure). A lightweight convolutional network is used to extract geometric features (such as the curvature of the mouth corners and the degree of eyelid opening) and dynamic features (such as the duration of micro-expressions) from the dynamic facial expression data in the unoccluded area. Then, a mapping relationship between muscle pressure features and facial expression features is established, generating an association weight matrix. Finally, a CNN-LSTM hybrid network is trained through supervised learning. The CNN branch processes the spatial features of the facial expression image (such as changes in periocular texture), while the LSTM branch analyzes the time series patterns of muscle pressure (such as the period of pressure fluctuation).

[0075] Training and Validation: Cross-validation: Use the Leave-One-Scenario-Out method to verify the model's generalization performance in unseen scenarios. Dynamic Optimization: Adjust model parameters based on real-time training feedback (such as user task completion) to avoid overfitting.

[0076] Output user-specific emotion classifiers to build a personalized psychological state model that can accurately reflect the user's psychological state.

[0077] When the user is undergoing VR training, dynamic data from the first acquisition unit and the second acquisition unit are acquired in real time, the dynamic data is input into the personalized psychological state model, the user's current psychological state is evaluated, and adjustment instructions are generated according to the evaluation results;

[0078] Specifically, the processing unit determines the user's current mental state through the following logic:

[0079] a. Data alignment and verification: comparing real-time facial muscle pressure change data with the preset pressure threshold range in the model. If the threshold is exceeded, an emotional abnormality flag is triggered;

[0080] b. Expression feature matching: extract key parameters from real-time facial expression dynamic data (such as the curvature of the mouth corners and the contraction frequency of the muscles around the eyes) and perform similarity matching with the emotion feature library in the model;

[0081] c. Comprehensive assessment decision-making, combining muscle pressure abnormality markers with expression matching results. If both point to the same emotional type (such as anxiety), it is determined to be that emotional state; if there is a conflict, priority is determined based on the preset weight coefficient (such as muscle pressure data weight accounts for 60%).

[0082] The adjustment unit is used to adjust the content presented in the VR glasses body in real time based on the adjustment instructions to match the user's current psychological state.

[0083] Specifically, generating adjustment instructions based on the evaluation results, the instructions include adjusting the difficulty of the virtual scene, switching the interactive plot type, or changing the intensity of physical interactive feedback;

[0084] The adjustment instructions adapt to the user's psychological state through the following logic:

[0085] If the patient is determined to be in an anxious state, the complexity of the scenario task is reduced, and guiding positive feedback (such as voice encouragement or virtual character assistance) is inserted;

[0086] If the player is judged to be in a focused state, the challenge difficulty will be gradually increased, and multi-threaded tasks will be added to enhance the player's ability to withstand stress;

[0087] If it is determined to be a relaxed state, the current scene parameters will be maintained and the user's immersive experience will be prolonged to consolidate the effect of cultivating psychological resilience.

[0088] Preferably, the device also includes a third acquisition unit and an optimization unit; the third acquisition unit consists of a high-precision heart rate sensor and a respiratory rate sensor. The heart rate sensor uses an optical transmission principle, emitting infrared light of a specific wavelength, penetrating the skin, and receiving reflected light to detect changes in blood volume, thereby accurately calculating the heart rate. The respiratory rate sensor uses capacitive sensing technology, which can sense the tiny changes in capacitance of the user's chest or abdomen caused by breathing, thereby accurately measuring the respiratory rate.

[0089] The core function of the optimization unit is to deeply integrate the heart rate and respiratory rate data obtained by the third acquisition unit with the evaluation results of the processing unit. Specifically, the optimization unit has a built-in data fusion algorithm. First, the heart rate and respiratory rate data are pre-processed, including filtering to remove noise, data smoothing, and time series alignment to ensure the accuracy and consistency of the data. Then, the processed physiological data is associated with the emotional state evaluation results output by the processing unit. For example, when the user is in an anxious state, the heart rate usually speeds up and the respiratory rate also increases accordingly. The optimization unit establishes a physiological-psychological state mapping model by analyzing the correlation pattern between this physiological data and the emotional state.

[0090] Based on the above mapping model, the optimization unit can adjust the scene adjustment strategy of the adjustment unit in real time. For example, if the user's breathing rate is irregular and high, and the heart rate exceeds the normal range, and the evaluation result of the processing unit has not yet been determined to be an anxiety state, the optimization unit will use this physiological feedback information as an auxiliary reference to send an early warning signal to the adjustment unit, appropriately reduce the difficulty of the scene, increase guiding feedback, and prevent the user from experiencing anxiety. In addition, when the user is in a focused state but the physiological data shows that the heart rate and breathing rate continue to rise, the optimization unit will dynamically adjust the rhythm of increasing the challenge difficulty to avoid excessive stimulation and causing psychological fatigue in the user. Through this dynamic optimization based on the biofeedback mechanism, the optimization unit enables the entire interactive toy to more accurately match the user's actual psychological state, improving the scientificity and effectiveness of psychological resilience training.

[0091] Preferably, a fourth acquisition unit and an abnormal data screening unit are also included; the fourth acquisition unit is composed of a high-precision vibration sensor. The sensor is based on the piezoelectric principle and can convert tiny vibrations into electrical signals. It has a wide-band response characteristic and can accurately capture vibration information within different frequency ranges. The vibration sensor is embedded on both sides of the VR glasses body near the user's temples. The bone structure here is relatively superficial, and when the user generates a vibration movement (such as coughing, head shaking when emotional, etc.), it can better transmit the vibration signal without causing any negative impact on the user's wearing comfort and visual experience.

[0092] The abnormal data screening unit is used to judge the user's current state based on the user's vibration frequency during VR training. If the vibration frequency exceeds a preset threshold, it is judged whether the user's current state is caused by emotional fluctuations. If so, the data collected by the first acquisition unit and the second acquisition unit are marked as valid data related to emotional fluctuations. Otherwise, they are marked as interference data.

[0093] Specifically, the abnormal data screening unit analyzes and determines the user's current status based on the following logic:

[0094] Set a preset threshold range for vibration frequency. This range is based on analysis of a large amount of preliminary experimental data and covers the vibration frequency range generated by daily user actions (such as normal speaking and slight head movements) under normal conditions.

[0095] When the vibration frequency collected by the fourth collection unit exceeds the preset threshold range, the abnormal state judgment process is triggered;

[0096] After the abnormality judgment process is triggered, the vibration frequency data is first analyzed in the time domain and frequency domain to extract the characteristic parameters of the vibration frequency, including but not limited to the amplitude of the vibration frequency (reflecting the vibration intensity), the frequency range (the vibration frequency range covered), the duration (the duration of the vibration), and the frequency change rate (the rate of change of the vibration frequency over time);

[0097] Retrieving a pre-built first cough-related characteristic parameter database, which stores vibration frequency characteristic parameters corresponding to different cough types (such as dry coughs, wet coughs, etc.) and intensities (such as light coughs and heavy coughs), the extracted vibration frequency characteristic parameters are compared with the vibration frequency characteristic parameters corresponding to coughs in the first characteristic parameter database, and the similarity between the two is calculated. If the similarity reaches or exceeds a preset cough matching threshold (such as 80%), the current vibration frequency phenomenon is determined to be caused by coughing, and the data currently collected by the first and second collection units are marked as interference data caused by coughing.

[0098] If the similarity with the cough characteristic parameter is lower than a preset cough matching threshold, a pre-established second characteristic parameter database related to emotional fluctuations is further retrieved. The second characteristic parameter database stores vibration frequency characteristic parameters corresponding to different emotional fluctuations (such as anxiety, anger, excitement, etc.). The extracted vibration frequency characteristic parameters are also compared with the vibration frequency characteristic parameters corresponding to emotional fluctuations in the second characteristic parameter database to calculate the similarity. If the similarity reaches or exceeds a preset emotional fluctuation matching threshold (for example, 75%), it is determined that the current vibration frequency phenomenon is caused by emotional fluctuations, and the data currently collected by the first and second collection units are marked as valid data related to emotional fluctuations, which is used for subsequent comprehensive assessment of the user's psychological state in combination with other collected data.

[0099] If the similarity with the cough and emotional fluctuation characteristic parameters does not reach the corresponding matching threshold, the data currently collected by the first acquisition unit and the second acquisition unit will be temporarily stored as pending data, and the changes in the vibration frequency data collected subsequently will continue to be observed. When the vibration frequency data collected subsequently are normal and the abnormal judgment process is not triggered again, the pending data will be regarded as accidental interference data and eliminated; if the abnormal judgment process is triggered again and the characteristic similarity with the pending data exceeds the preset similarity threshold (for example, 60%), the two will be combined and the above-mentioned comparison analysis will be performed again to dynamically optimize the judgment results and ensure the accuracy of data labeling, thereby providing a reliable data basis for psychological state assessment.

[0100] Specifically, the following method can be used to calculate the similarity between the vibration frequency characteristic parameters and the vibration frequency characteristic parameters corresponding to coughs in the first characteristic parameter database (or emotional fluctuations in the second characteristic parameter database). In this embodiment, the similarity between the vibration frequency characteristic parameters corresponding to coughs in the first characteristic parameter database is calculated as an example:

[0101] Feature parameter preprocessing and standardization

[0102] Parameter extraction range:

[0103] Amplitude: Peak value of vibration acceleration (0.1-10g), reflecting the intensity of cough (e.g., 0.5-1.2g for mild cough, 2-5g for severe cough);

[0104] Frequency range: distribution of the main frequency band (typical cough frequency range is 5-15 Hz), calculation of energy concentration (e.g. > 60% of the energy is in this frequency band);

[0105] Duration: effective vibration duration (cough pulse width 0.1-0.5 seconds);

[0106] Frequency change rate: The frequency slope of the vibration signal in the time domain (coughing is manifested as a sudden high-frequency shock with a change rate ≥50 Hz / s).

[0107] Normalization processing:

[0108] Amplitude: linearly mapped to [0,1] according to the sensor range (0-10g);

[0109] Frequency range: Calculate the proportion of each sub-band based on the preset maximum bandwidth (e.g. 0-50Hz);

[0110] Duration: The denominator is the longest allowed event duration (e.g., 3 seconds), normalized to [0, 1];

[0111] Frequency change rate: Scale the frequency based on the maximum change rate calculated from historical data (e.g., 200 Hz / s).

[0112] Single dimension similarity calculation:

[0113]

[0114] Where A1 is the amplitude of the real-time vibration signal (unit: g), that is, the peak value of the vibration acceleration currently collected (such as 2.3g measured when coughing); A2 is the amplitude of the cough feature sample stored in the database (such as the typical amplitude of the standard cough sample is 1.8g); A range This is the maximum amplitude of the sensor's range (e.g., 0-10g). It is used for normalization. By comparing the absolute difference between the real-time amplitude and the database samples, the normalized value reflects the degree of similarity (the closer the value is to 1, the more similar the amplitudes are).

[0115] F1 is the frequency band energy distribution vector (dimension: n) of the real-time vibration signal, which is composed of the energy proportion of each sub-band (such as 5 Hz as a frequency band); F2 is the frequency band energy distribution vector of the cough feature sample in the database; it measures the directional consistency of the two frequency band energy vectors (the closer the value is to 1, the more similar the frequency distribution is).

[0116] T1 is the effective duration of the real-time vibration signal (unit: seconds), that is, the time from the vibration amplitude exceeding the threshold to the return to calm (such as a cough lasting 0.3 seconds); T2 is the standard duration of the cough feature sample in the database (such as a typical cough lasting 0.25 seconds); the duration matching degree is quantified by the time overlap ratio (the closer the value is to 1, the more consistent the duration).

[0117] D1 is the optimal path distance calculated by the dynamic time warping (DTW) algorithm, which reflects the alignment cost of the real-time signal and the database sample in terms of frequency change trend; D max is the preset maximum allowed distance threshold; convert the DTW distance into a similarity score (the smaller the distance, the higher the similarity).

[0118] Weight Distribution: Example

[0119] characteristic parameters Weight Frequency range 35％ Amplitude 25％ Frequency change rate 30％ Duration 10％

[0120] Total similarity calculation:

[0121] S 总 =0.35*S2+0.25*S1+0.3*S4+0.1*S3

[0122] Preferably, the system further includes a fifth acquisition unit and a verification unit; the fifth acquisition unit is configured to collect user voice data. In this embodiment, the fifth acquisition unit is configured with a microphone array integrated on the outside of the VR glasses body to directionally pick up the user's voice. The voice data includes voice content, pitch frequency, and non-speech acoustic events (such as coughing and sobbing).

[0123] The verification unit is used to verify the accuracy of the judgment of the interference data based on the sound data.

[0124] Specifically, the characteristic parameters of the sound data collected by the fifth acquisition unit are extracted. The characteristic parameters include, but are not limited to, the frequency range (such as the frequency of a typical cough sound), the intensity range (such as sound pressure level), the duration, and the specific shape of the sound waveform by performing a short-time Fourier transform (STFT) on the sound data. The characteristic parameters of the sound data are compared with the characteristic parameters in a pre-established cough sound feature database, and the similarity between the two in terms of frequency, intensity, duration, and waveform is calculated.

[0125] Specifically, feature parameter preprocessing and standardization

[0126] Feature extraction and normalization:

[0127] Sound characteristics:

[0128] Frequency range: Extract the energy proportion of the main frequency band (such as 200-800Hz of cough sound) and normalize it to [0,1];

[0129] Intensity range: Calculate the root mean square (RMS) sound pressure level (SPL) and linearly map it to [0,1] according to the maximum dynamic range (e.g. 0-100dB);

[0130] Duration: Record the start and end time of the acoustic event (e.g., cough pulse width 0.2 seconds), and normalize it to the preset maximum duration (e.g., 3 seconds);

[0131] Waveform morphology: Extract the short-term zero-crossing rate (ZCR) and Mel-frequency cepstral coefficients (MFCC 1-12), and form a feature vector after dimensionality reduction through principal component analysis (PCA).

[0132] Vibration characteristics:

[0133] Amplitude: peak value of vibration acceleration (e.g., 0.5-5g), normalized to [0,1];

[0134] Frequency range: Main frequency band (e.g. cough vibration 5-15Hz vs. emotional fluctuation 3-8Hz), calculate the similarity of frequency domain energy distribution;

[0135] Duration: The effective duration of the vibration signal, normalized to the preset maximum duration (such as 10 seconds).

[0136] Time window alignment:

[0137] Based on the IEEE 1588 clock synchronization protocol, the sound and vibration data are aligned within a ±50ms time window to ensure that the characteristic parameters correspond to the same event.

[0138] Single dimension similarity calculation:

[0139] Cosine similarity:

[0140]

[0141] Where, is the real-time data feature vector, is the database feature vector.

[0142] Multi-feature weighted fusion:

[0143] Weight distribution (taking cough matching as an example):

[0144]

[0145]

[0146] Comprehensive similarity calculation:

[0147]

[0148] Where w i is the weight of each feature, Similarity i is the normalized single-dimensional similarity.

[0149] If the similarity between the sound data and the cough sound characteristics reaches or exceeds the preset cough sound matching threshold, it is determined that the current vibration frequency phenomenon is caused by coughing, and the relevant data is confirmed to be interference data; otherwise, the interference data is changed to valid data.

[0150] The above is only an embodiment of the present invention, and common knowledge such as the specific structure and / or characteristics of the scheme is not described in detail here. It should be pointed out that for those skilled in the art, without departing from the structure of the present invention, several variations and improvements can be made, which should also be regarded as the scope of protection of the present invention, and these will not affect the effect of the implementation of the present invention and the practicality of the patent. The scope of protection required by this application shall be based on the content of its claims, and the specific implementation methods and other records in the specification can be used to interpret the content of the claims.

Claims

1. An interactive toy for cultivating psychological resilience in adolescents, characterized by: It comprises a VR glasses body (1), a first acquisition unit, a second acquisition unit, a processing unit and an adjustment unit; The VR glasses body (1) is used to provide a virtual reality training scene for the user; The first acquisition unit is integrated into the VR glasses body (1), and is used to collect facial muscle pressure change data of the user in the area blocked by the VR glasses; The second acquisition unit is used to synchronously acquire dynamic facial expression data of the user's unobstructed area; The processing unit is configured to establish a personalized psychological state model of the user based on the correlation between the muscle pressure data of the occluded area and the facial expression data of the unoccluded area; and when the user is performing VR training, acquire dynamic data from the first acquisition unit and the second acquisition unit in real time, input the dynamic data into the personalized psychological state model, evaluate the user's current psychological state, and generate adjustment instructions based on the evaluation results; The adjustment unit is used to adjust the content presented in the VR glasses body (1) in real time based on the adjustment instruction to match the user's current psychological state.

2. The interactive toy for cultivating psychological resilience in adolescents according to claim 1, characterized in that: The processing unit determines the user's current mental state through the following logic: a. Data alignment and verification: comparing real-time facial muscle pressure change data with the preset pressure threshold range in the model. If the threshold is exceeded, an emotional abnormality flag is triggered; b. Expression feature matching: extract key parameters from real-time facial expression dynamic data and perform similarity matching with the emotion feature library in the model; c. Comprehensive assessment decision: combining abnormal muscle pressure markers with expression matching results. If both indicate the same emotional type, then the emotional state is determined to be that type. If there is a conflict, the priority is determined according to the preset weight coefficient.

3. The interactive toy for cultivating psychological resilience in adolescents according to claim 2, characterized in that: The adjustment instructions adapt to the user's psychological state through the following logic: If it is determined to be an anxious state, the complexity of the scene task is reduced and guiding positive feedback is inserted; If the player is judged to be in a focused state, the challenge difficulty will be gradually increased, and multi-threaded tasks will be added to enhance the player's ability to withstand stress; If it is determined to be a relaxed state, the current scene parameters will be maintained and the user's immersive experience will be prolonged to consolidate the effect of cultivating psychological resilience.

4. The interactive toy for cultivating psychological resilience in adolescents according to claim 3, characterized in that: The steps of constructing the personalized mental state model include: S1: Collect user baseline data through a preset standardized emotion induction experiment; S2: Perform time-frequency domain analysis on the facial muscle pressure change data in the occluded area to extract the pressure distribution pattern and dynamic change gradient features; extract geometric features and dynamic features of the facial expression dynamic data in the unoccluded area through a lightweight convolutional network; S3: Establish the mapping relationship between muscle pressure features and facial expression features, and generate the correlation weight matrix; S4: Train the CNN-LSTM hybrid network through supervised learning and output a user-specific emotion classifier.

5. The interactive toy for cultivating psychological resilience in adolescents according to claim 4, characterized in that: It also includes a third acquisition unit and an optimization unit; the third acquisition unit is used to collect the user's heart rate and respiratory rate data; the optimization unit is used to jointly analyze the heart rate and respiratory rate data with the evaluation results of the processing unit based on the biofeedback mechanism, and dynamically optimize the scene adjustment strategy.

6. The interactive toy for cultivating psychological resilience in adolescents according to claim 5, characterized in that: The first acquisition unit comprises a plurality of flexible piezoresistive sensor arrays embedded in the inner contact surface of the VR glasses body (1) in a matrix form.

7. The interactive toy for cultivating psychological resilience in adolescents according to claim 6, characterized in that: The first acquisition unit covers the user's cheekbone, glabella, nose bridge, frontalis muscle and temporalis muscle area.

8. The interactive toy for cultivating psychological resilience in adolescents according to claim 7, characterized in that: The second acquisition unit is a camera, which is arranged on the side of the VR glasses body close to the user's unobstructed face.

9. The interactive toy for cultivating psychological resilience in adolescents according to claim 8, characterized in that: It also includes a feedback unit; the feedback unit is used to output corresponding encouragement or guidance information to the user based on the user's performance in VR training and the psychological state evaluation results.

10. The interactive toy for cultivating psychological resilience in adolescents according to claim 9, characterized in that: It also includes a fourth acquisition unit and an abnormal data screening unit; the fourth acquisition unit is used to collect the vibration frequency of the user during VR training; the abnormal data screening unit is used to judge the user's current state based on the user's vibration frequency during VR training. If the vibration frequency exceeds a preset threshold, it is judged whether the user's current state is caused by emotional fluctuations. If so, the data collected by the first acquisition unit and the second acquisition unit are marked as valid data related to emotional fluctuations, otherwise, they are marked as interference data.

11. The interactive toy for cultivating psychological resilience in adolescents according to claim 10, characterized in that: It also includes a fifth collection unit and a verification unit; the fifth collection unit is used to collect sound data emitted by the user; the verification unit is used to verify the accuracy of the interference data judgment based on the sound data.

12. The interactive toy for cultivating psychological resilience in adolescents according to claim 11, characterized in that: The abnormal data screening unit analyzes and judges the user's current status based on the following logic: Setting a preset threshold range for the vibration frequency, and triggering an abnormal state judgment process when the vibration frequency collected by the fourth collection unit exceeds the preset threshold range; After the abnormality judgment process is triggered, the vibration frequency data is first analyzed in the time domain and frequency domain to extract the characteristic parameters of the vibration frequency, including but not limited to the amplitude, frequency range, duration and frequency change rate of the vibration frequency; A first characteristic parameter database related to cough constructed in advance is retrieved. The first characteristic parameter database stores vibration frequency characteristic parameters corresponding to coughs of different types and intensities. The extracted vibration frequency characteristic parameters are compared with the vibration frequency characteristic parameters corresponding to coughs in the first characteristic parameter database, and the similarity between the two is calculated. If the similarity reaches or exceeds a preset cough matching threshold, it is determined that the current vibration frequency phenomenon is caused by coughing, and the data currently collected by the first acquisition unit and the second acquisition unit are marked as interference data caused by coughing.

13. The interactive toy for cultivating psychological resilience in adolescents according to claim 12, characterized in that: The verification unit performs verification based on the following logic: extracting characteristic parameters of the sound data collected by the fifth collection unit; Comparing the characteristic parameters of the sound data with the characteristic parameters in a cough sound characteristic database constructed in advance; If the similarity between the sound data and the cough sound characteristics reaches or exceeds the preset cough sound matching threshold, it is determined that the current vibration frequency phenomenon is caused by coughing, and the relevant data is confirmed to be interference data; otherwise, the interference data is changed to valid data.

Images