Multifunctional seat system with immersive experience function for psychological experiment and evaluation

Abstract

The invention provides a multifunctional seat system with an immersive experience function for psychological experiments and evaluation, and relates to the technical field of man-machine interaction. The system comprises a heterogeneous sensing module which collects physiological and mechanical data in real time and converts the physiological and mechanical data into a pressure gradient tensor, a dynamic actuating module adjusts the shape and inertial disturbance of a seat through a shape memory alloy and an electromagnetic damping base, and an environment evolution module maps a pressure extreme value into a virtual curvature singular point and reconstructs a scene physical rule. The cognitive coupling module constructs a psychological state transition model based on the physiological and behavior data and outputs a predicted value; the four-dimensional equilibrium state detection realizes a multi-dimensional constraint closed loop, and triggers a data flow path inversion protocol and dynamic parameter optimization; through real-time correction of a geodesic equation, a discretization Ricci flow algorithm and a gradient back projection technology, seat dynamic response and virtual scene evolution are synchronized, and psychological state prediction, inertial disturbance intensity adaptive adjustment and multi-dimensional coupling of an immersive experimental environment are realized.

Description

Technical Field

[0001] The present invention relates to the field of human-computer interaction technology, and in particular to a multifunctional chair system for psychological experiments and assessments with an immersive experience function. Background Art

[0002] Psychological experiments and assessment technologies are of great value in modern psychological research and clinical diagnosis. With the integration of virtual reality and biofeedback technology, immersive experimental environments can enhance subject participation and the authenticity of data collection. Current psychological research has an increasing demand for dynamic interactive scenarios, and there is an urgent need for systems that can synchronously perceive physiological and behavioral data and adjust the physical environment in real time to more accurately simulate real psychological scenarios.

[0003] In the existing technology, psychological assessment equipment mostly uses a combination of fixed seats and basic sensors; for example, heart rate or electromyography signals are collected through a single modality sensor, and visual stimulation is provided in conjunction with static virtual reality images; some equipment integrates a simple motion platform to simulate physical disturbances through preset vibration patterns.

[0004] The existing technology has the following shortcomings: it is difficult for the laboratory environment to simulate the stress sources of the real scene, The lack of a real-time physiological and behavioral feedback loop makes it impossible to dynamically adjust the stimulation intensity according to the subject's anxiety level to provide comfort; physiological signals and environmental parameters are not synchronously coupled, which reduces the accuracy of state assessment. Summary of the Invention

[0005] (1) Technical problems solved In response to the shortcomings of the existing technology, the present invention provides a multifunctional chair system for psychological experiments and assessments with immersive experience functions, so as to solve the problems raised in the above background technology, namely, that the laboratory environment is difficult to simulate the stress sources of real scenes, lacks a real-time physiological and behavioral feedback loop, and cannot dynamically adjust the stimulation intensity according to the anxiety level of the subject to provide comfort; and the physiological signals and environmental parameters are not synchronously coupled, which reduces the accuracy of state assessment.

[0006] (2) Technical solution To achieve the above objectives, the present invention is implemented through the following technical solutions: a multifunctional seat system for psychological experiments and assessments with immersive experience functions, including a heterogeneous sensing module, a dynamic actuation module, an environmental evolution module, a cognitive coupling module, and a four-dimensional equilibrium state detection module; The heterogeneous sensing module includes a physiological signal acquisition unit and a metabolic parameter sensor array, which are used to synchronously capture heart rate variability, skin conductance response, surface electromyographic signals, and body surface temperature gradient data. It also uses a distributed pressure sensing array to capture the subject's body pressure distribution and micro-motion trajectory in real time, and converts the original mechanical signal into a pressure gradient tensor. When it is detected that the angle between the pressure gradient direction vector and the deformation direction vector of the dynamic actuation module reaches a preset threshold through vector dot product calculation, a data flow path reversal protocol is triggered to force the pressure gradient tensor to be transmitted to the environmental evolution module for preprocessing. The environmental evolution module maps the extreme points of the pressure gradient tensor to curvature singularities of the virtual space, and modifies the geodesic equation based on the deformation variable fed back by the dynamic actuation module to reconstruct the physical laws of the scene. Simultaneously, the inverse vector of the psychological state prediction value of the previous frame output by the cognitive coupling module is embedded as a curvature distortion factor. The inverse vector is generated by the cognitive coupling module through a gradient backprojection algorithm. Based on the modified geodesic equation and the curvature distortion factor, the second-order derivative of the Gaussian curvature of the virtual space is solved through a discretized Ricci flow algorithm to generate a curvature derivative instruction. The dynamic actuation module receives the curvature derivative instruction output by the environmental evolution module, dynamically adjusts the curvature of the supporting surface through the shape memory alloy actuator, and drives the electromagnetic damping motion base to generate inertial perturbations that match the physical laws of the virtual scene. Simultaneously, it receives the psychological state prediction value output by the cognitive coupling module, encodes it into the actuator deformation rate and the electromagnetic damping motion base acceleration, and combines it with the curvature derivative instruction to optimize the seat shape and movement pattern in real time, so that the inertial perturbation intensity generated by the seat is proportional to the subject's estimated psychological energy value. The cognitive coupling module generates a psychological energy estimation value by real-time analysis of the spectrum entropy of physiological signals and the mutation frequency of behavioral data, combined with the prediction results of the psychological state migration direction by the Lyapunov index. At the same time, a psychological state migration model is constructed through a dynamic Bayesian network to generate a psychological state prediction value. The psychological state prediction value is encoded as the motion parameter of the dynamic actuation module and the input parameter of the curvature distortion factor of the environmental evolution module. The curvature distortion factor of the environmental evolution module interacts with the temporal misalignment of the cognitive coupling module to form a recursive constraint among the four modules. The heterogeneous sensing module, dynamic actuation module, environmental evolution module, and cognitive coupling module achieve closed-loop coupling of data streams and control signals through four-dimensional equilibrium state detection. The four-dimensional equilibrium state detection includes a multi-dimensional constraint relationship between the extreme point distribution of the heterogeneous sensing module, the deformation derivative of the dynamic actuation module, the curvature distortion factor of the environmental evolution module, and the psychological energy estimation value of the cognitive coupling module. Only when all the constraint conditions are synchronously met is the system allowed to output a valid control signal.

[0007] Preferably, the heterogeneous sensing module captures the subject's heart rate variability, skin conductance response, surface electromyography signal, and body surface temperature gradient data in real time through a physiological signal acquisition unit; the metabolic parameter sensing array synchronously collects metabolic parameters and captures the subject's body pressure distribution and micro-motion trajectory through a distributed pressure sensing array at a preset sampling rate; the distributed pressure sensing array converts discrete pressure values ​​into a pressure gradient tensor in a continuous space-time field. The specific operation is as follows: within each sampling period, the spatial coordinates of the pressure sensing nodes are modeled on a Riemannian manifold, and the covariant derivatives of the pressure values ​​of each node are calculated to extract the saddle point and extreme point distribution characteristics; when it is detected that the angle between the pressure gradient direction vector and the deformation direction vector of the dynamic actuation module reaches a preset threshold through vector dot product calculation, a data flow path reversal protocol is triggered, and the pressure gradient tensor is transmitted to the environmental evolution module for preprocessing; the execution steps of the data flow path reversal protocol include: interrupting the current data flow transmission path, re-encoding the original mechanical signal into a tensor format, and reallocating the transmission channel to the curvature singularity mapping unit of the environmental evolution module through a priority arbitration mechanism.

[0008] Preferably, after receiving the pressure gradient tensor transmitted by the heterogeneous sensing module, the environmental evolution module maps its extreme points to curvature singularities of the virtual space; the curvature singularity is a high curvature area generated by mapping the extreme points of the pressure gradient tensor in the virtual space, and its curvature value is higher than 30% of the surrounding environment, which is used to simulate the local perturbation effect of body pressure concentration / release on the physical laws of the scene; wherein the extreme point is defined as the local maximum or minimum value of the pressure gradient tensor in the space-time field, and its physical meaning represents the concentration or release area of ​​the subject's body pressure distribution. For example, when the subject's psychological state changes, such as tension or relaxation, resulting in local concentration of body pressure distribution, the pressure gradient tensor will present a maximum value in the corresponding area; conversely, when the body pressure is released, such as movement adjustment, a minimum value is formed, and its mathematical expression is: Where P(x, y, t) is the pressure distribution function in the space-time field. The detected extreme point coordinates are mapped one-to-one to the curvature singularity of the virtual space. The virtual curvature parameters are adjusted to simulate the impact of body pressure concentration or release on the physical laws of the scene. The mapping process solves the second-order derivative of the Gaussian curvature of the virtual space through the discretized Ricci flow algorithm to generate the curvature derivative instruction. The formula is: , where g ij is the metric tensor of the virtual space, R ij is the Ricci curvature tensor, f is the correction term of the geodesic equation, δ ij is the Kronecker symbol; the curvature distortion factor is the inverse vector V output by the cognitive coupling module reverse The product of the preset proportional coefficient k is generated, that is, , where the inverse vector ;in, is the loss function of the psychological state prediction value, P is the psychological state transition probability distribution output by the dynamic Bayesian network, and η is the learning rate; the specific operation is: based on the real-time deformation variable feedback from the dynamic actuation module, the geometric parameter terms of the geodesic equation, including the gravity field intensity distribution and the material viscoelastic coefficient, are modified; at the same time, the inverse vector of the psychological state prediction value of the previous frame output by the cognitive coupling module is embedded into the geodesic equation as a curvature distortion factor; the inverse vector is generated by the gradient backprojection algorithm, and its value is opposite to the gradient direction of the psychological state prediction value; the modified geodesic equation is discretized by the finite difference method, and then the curvature field evolution of the virtual space is solved in combination with the curvature distortion factor, and finally the curvature derivative instruction for driving the dynamic actuation module is generated.

[0009] Preferably, the dynamic actuation module receives the curvature derivative instruction output by the environment evolution module, analyzes its sign and amplitude, and dynamically sets the deformation rate threshold of the shape memory alloy actuator; the deformation rate threshold is adjusted in real time according to the residual term of the geodesic equation solved by the environment evolution module, and the calculation formula of the residual term is the absolute value of the difference between the current deformation variable and the theoretical deformation variable. When the residual term exceeds 0.1, the deformation rate threshold is decreased by a proportional coefficient of 0.8; if the residual term is lower than 0.05, the threshold is increased by a coefficient of 1.2; the adjustment process is implemented by a PID controller to ensure the dynamic matching of the deformation rate and the physical law of the virtual scene; at the same time, the output of the cognitive coupling module is The psychological state prediction value is parsed into a deformation rate weight factor and inertial disturbance waveform parameters; the deformation rate weight factor corrects the actual deformation of the actuator through a nonlinear mapping rule, and optimizes the current deformation path in combination with the deformation parameters stored in the historical deformation data buffer; the inertial disturbance waveform parameters generate a non-periodic vibration waveform by decomposing the psychological state prediction value into amplitude, frequency and phase components, where the amplitude is proportional to the psychological energy estimate, the frequency component controls the random distribution of the vibration interval, and the phase component is generated based on a chaotic phase superposition algorithm of logistic mapping; finally, an inertial disturbance matching the physical laws of the virtual scene is output through an electromagnetic damping motion base.

[0010] Preferably, the cognitive coupling module analyzes the spectrum entropy of the physiological signal and the mutation frequency of the behavioral data collected by the heterogeneous sensing module in real time to extract a multidimensional feature vector; predicts the direction of mental state migration through the Lyapunov index and corrects the weight distribution of the feature vector; the dynamic Bayesian network generates a mental state migration probability distribution based on the pre-trained mental state database, and outputs a maximum likelihood estimate as the mental state prediction value of the current frame; the step of generating the predicted value includes: performing wavelet denoising and normalization processing on the physiological signal, extracting time domain and frequency domain features, inputting the dynamic Bayesian network to calculate the migration probability, and selecting the category with the highest probability as the predicted value; the predicted value generates an inverse vector through the gradient back projection algorithm, which is used as the curvature distortion factor input parameter of the environmental evolution module; at the same time, by adjusting the ratio of the sampling clock frequency of the heterogeneous sensing module to the rendering frame rate to an incommensurable transcendental number, periodic interference is avoided, and the transcendental number uses π as the default value; the Lyapunov index λ is used to quantify the degree of chaos of the mental state migration model, and its calculation is based on the physiological signal time series {x t}: , where f(x) is the nonlinear mapping function of the psychological state migration direction; the physiological signal spectrum entropy is obtained in real time by the physiological signal acquisition unit of the heterogeneous sensing module. After wavelet denoising and normalization processing, the signal spectrum characteristics are extracted using frequency domain transformation, and the complexity of the signal is reflected by entropy calculation; the behavioral data mutation frequency is captured by a distributed pressure sensor array. The body pressure distribution and micro-movement trajectory data of the subject are captured, the time series signal is differentiated, the mutation points in the pressure gradient or movement trajectory are identified, the frequency of mutation events per unit time is counted, and the mutation judgment conditions are calibrated in combination with the deformation feedback data of the dynamic actuation module. Finally, the quantized mutation frequency parameters are output.

[0011] Preferably, the four-dimensional equilibrium state detection module realizes closed-loop coupling of data flow and control signal by verifying the multi-dimensional constraint relationship between the extreme point distribution of the heterogeneous sensing module, the deformation derivative of the dynamic actuation module, the curvature distortion factor of the environmental evolution module and the psychological energy estimation value of the cognitive coupling module; the specific conditions include: the extreme point distribution of the heterogeneous sensing module and the deformation derivative of the dynamic actuation module satisfy the Hodge duality relationship, that is, the tensor outer product of the two is zero; the absolute value of the curvature distortion factor of the environmental evolution module is equal to the product of the psychological energy estimation value and the preset proportional coefficient; the ratio of the sampling rate of the heterogeneous sensing module to the rendering frame rate of the environmental evolution module is a transcendental number; only when all the constraints are met synchronously, the system allows the output of a valid control signal; if any condition deviates from the threshold, the four-dimensional equilibrium state reconstruction process is immediately triggered, including resetting the curvature distortion factor, rolling back the deformation parameters to the history buffer, and reinitializing the data flow path.

[0012] Preferably, the data flow path reversal protocol is triggered when the heterogeneous sensing module detects that the angle between the pressure gradient direction and the deformation direction vector exceeds a preset threshold, and the default value of the preset threshold is 30 degrees; the protocol execution steps include: pausing the current data flow transmission, and re-encoding the pressure gradient tensor into a space-time continuous field format; allocating the transmission channel to the curvature singularity mapping unit of the environmental evolution module through a priority arbitration mechanism; the encoding process includes calculating the Riemannian manifold covariant derivative of the pressure gradient tensor and extracting the distribution characteristics of saddle points and extreme points; at the same time, the protocol forces the data flow to bypass the default path and transmit it directly to the environmental evolution module for preprocessing; after the preprocessing is completed, the data flow re-enters the deformation control unit of the dynamic actuation module to ensure a real-time feedback closed loop.

[0013] Preferably, the deformation rate threshold of the shape memory alloy actuator is dynamically set according to the curvature derivative instruction output by the environmental evolution module; the specific operation is: analyzing the sign and amplitude of the curvature derivative instruction, if the derivative is positive, the deformation rate threshold is increased, otherwise it is reduced; at the same time, combined with the deformation rate weight factor output by the cognitive coupling module, the actual deformation amount is corrected through nonlinear mapping rules; the weight factor is dynamically adjusted according to the timing characteristics of the psychological state prediction value; the historical deformation data buffer stores the deformation parameters of the last N cycles for optimizing the current deformation path; when the psychological state prediction value exceeds the safety threshold, the priority arbitration mechanism forcibly suppresses the deformation rate and acceleration to ensure the safety of the subject. The safety threshold is the upper limit threshold of the psychological state prediction value, and the default setting is 0.8.

[0014] Preferably, the environmental evolution module drives the evolution of the physical laws of the virtual scene by modifying the geodesic equation, including the distribution of gravity field intensity and the viscoelastic coefficient of the material; the modification of the geodesic equation is based on the real-time deformation variable feedback from the dynamic actuation module, so that the motion trajectory of the virtual object presents a nonlinear path deviation related to the deformation variable; the deviation amplitude is positively correlated with the real-time deformation variable; at the same time, the inertial disturbance waveform parameters generated by the electromagnetic damping motion base extract the virtual scene inertia parameters from the curvature derivative instructions, and dynamically adjust the energy distribution of the vibration waveform; the actual acceleration data is fed back to the cognitive coupling module for calibrating the psychological state prediction model to ensure a high degree of synchronization between the virtual scene and the seat movement.

[0015] Preferably, the Lyapunov index is used to evaluate the stability of the psychological state transition model and predict the migration direction of the subject's psychological state; the specific operation is: real-time calculation of the Lyapunov index of the physiological signal spectrum entropy and the behavioral data mutation frequency, if the index exceeds the preset threshold, the parameter recalibration process of the dynamic Bayesian network is triggered; the index is also used to correct the weight distribution of the multidimensional feature vector and optimize the generation of the psychological state prediction value; at the same time, the Lyapunov index is dynamically bound to the sampling clock frequency of the heterogeneous sensing module to ensure that the ratio of the sampling rate to the rendering frame rate is a transcendental number, thereby avoiding the influence of periodic interference on the prediction results.

[0016] Preferably, when the four-dimensional equilibrium state detection module detects that any constraint condition deviates from the threshold, the reconstruction process is triggered immediately; the priority is as follows: resetting the curvature distortion factor generation rule of the environmental evolution module to the initial proportional coefficient; rolling back the deformation derivative of the dynamic actuation module to the last valid state, calling the historical deformation data buffer to restore the parameters; forcing the heterogeneous sensing module to reinitialize the data flow path and re-execute the data flow path reversal protocol; the rollback operation overwrites the current state by reading the most recent valid deformation parameter in the buffer; after the reconstruction is completed, the system re-verifies the multi-dimensional constraint relationship until all conditions are met and the control signal output is restored.

[0017] Preferably, when the Lyapunov exponent exceeds a preset threshold, the dynamic Bayesian network triggers a parameter recalibration process; the specific steps include: pausing the current mental state prediction task; obtaining the latest physiological signals and behavioral data from the heterogeneous sensing module, and re-extracting the multidimensional feature vector; adjusting the feature weight distribution in combination with the updated Lyapunov exponent; retraining the dynamic Bayesian network using the standard mental state database; after the calibration is completed, resuming the prediction task and outputting the corrected mental state prediction value; the calibration process ensures the convergence of the model and avoids the deviation of the prediction value due to system disturbance.

[0018] Preferably, the phase component of the non-periodic vibration waveform is generated by a chaotic phase superposition algorithm of a Logistic map; the specific operations are: inputting the phase component of the psychological state prediction value into the Logistic map equation to generate a chaotic sequence; converting the chaotic sequence into a phase offset through normalization processing; superimposing the offset on the basic vibration waveform to generate a non-periodic vibration signal; at the same time, combining the virtual scene inertia parameters in the curvature derivative instruction, dynamically adjusting the energy distribution of the waveform; finally, outputting an inertial disturbance that matches the virtual physical law through the electromagnetic damping motion base, and its amplitude is strictly proportional to the psychological energy estimation value.

[0019] (3) Beneficial effects The present invention provides a multifunctional seat system for psychological experiments and assessments with immersive experience functions. It has the following beneficial effects: 1. The present invention uses a heterogeneous sensing module to capture physiological and mechanical data in real time and converts it into a pressure gradient tensor. Combined with the curvature singularity mapping and geodesic equation correction of the environmental evolution module, the dynamic actuation module generates inertial perturbations that match the physical laws of the virtual scene. Multidimensional data flow closure and four-dimensional equilibrium state detection ensure the real-time and stability of the system response. The cognitive coupling module accurately predicts psychological states based on the Lyapunov exponent and dynamic Bayesian network, optimizing the seat shape and perturbation intensity. The non-periodic vibration waveform and chaotic phase superposition algorithm enhance the realism of the immersive experience. The system synchronously triggers safe reconstruction through multi-dimensional constraints to ensure the safety of the subjects and the reliability of the evaluation.

[0020] 2. The real-time interaction between the heterogeneous sensing module and the dynamic actuation module of the present invention enables adaptive adjustment of the seat shape. The environmental evolution module dynamically corrects the physical laws of the virtual scene through the discretized Ricci flow algorithm. The cognitive coupling module extracts feature vectors using spectral entropy and behavioral mutation frequency, and outputs high-precision psychological state prediction values ​​in combination with a dynamic Bayesian network. The four-dimensional equilibrium state detection forcibly constrains the data flow and control signals to avoid periodic interference and ensure system convergence. The historical deformation buffer and parameter recalibration process enhance the system's fault tolerance. The chaotic phase vibration waveform generated by the logistic map is strictly synchronized with the psychological energy estimate, improving the objectivity and scene adaptability of the evaluation results. DETAILED DESCRIPTION

[0021] The following is a clear and complete description of the technical solutions in the embodiments of the present invention. Obviously, the embodiments described are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making any creative efforts are within the scope of protection of the present invention.

[0022] An embodiment of the present invention provides a multifunctional chair system for psychological experiments and assessments with an immersive experience function, which includes: after the subject enters the experimental environment, the system starts an initialization process, and first the physiological signal acquisition unit and the metabolic parameter sensor array of the heterogeneous sensing module begin to synchronously capture the subject's heart rate variability, skin conductance response, surface electromyography signal and body surface temperature gradient data; the distributed pressure sensing array monitors the subject's body pressure distribution and micro-movement trajectory in real time at a sampling rate of 1000 times per second, and converts the discrete pressure value into a pressure gradient tensor in a continuous space-time field through Riemannian manifold modeling.

[0023] When it is detected that the angle between the pressure gradient direction and the deformation direction vector of the dynamic actuation module reaches a preset threshold of 60 degrees, the data flow path reversal protocol is triggered, the default data transmission path is suspended, the pressure gradient tensor is re-encoded into a space-time continuous field format, and allocated to the curvature singularity mapping unit of the environmental evolution module through a priority arbitration mechanism; at this time, after receiving the pressure gradient tensor, the environmental evolution module uses the discretized Ricci flow algorithm to extract its extreme points, maps them to the curvature singularities of the virtual space, and corrects the geometric parameter terms of the geodesic equation based on the real-time deformation feedback from the dynamic actuation module, including the gravity field intensity distribution and the material viscoelastic coefficient.

[0024] At the same time, the cognitive coupling module analyzes the spectrum entropy of physiological signals and the mutation frequency of behavioral data, combines the Lyapunov exponent to predict the direction of psychological state migration, generates a psychological state prediction value for the current frame, and generates an inverse vector through the gradient backprojection algorithm. This vector is embedded into the geodesic equation as a curvature distortion factor. The modified geodesic equation is discretized and solved using the finite difference method to generate the second-order derivative of the Gaussian curvature of the virtual space, forming a curvature derivative instruction and transmitting it to the dynamic actuation module. The dynamic actuation module analyzes the sign and amplitude of the curvature derivative instruction and dynamically sets the deformation rate threshold of the shape memory alloy actuator. If the derivative is positive, the threshold is increased to 0.5 mm / s, otherwise it is reduced to 0.3 mm / s. At the same time, the psychological state prediction value output by the cognitive coupling module is analyzed into a deformation rate weight factor and inertial perturbation waveform parameters. The deformation rate weight factor modifies the actual deformation variable through a nonlinear mapping rule. For example, when the predicted value is a high anxiety state, the weight factor is increased to 1.2 times. The historical deformation data buffer stores the most recent 10 The deformation parameters of each cycle are used to optimize the current deformation path; the inertial disturbance waveform parameters are decomposed into amplitude, frequency and phase components by decomposing the psychological state prediction value, wherein the amplitude is positively correlated with the psychological energy estimation value according to a proportional coefficient of 0.8, the frequency component controls the random distribution of the vibration interval, and the phase component generates a non-periodic vibration waveform based on the chaotic phase superposition algorithm of the Logistic map; the electromagnetic damping motion base is combined with the virtual scene inertia parameters in the curvature derivative instruction to dynamically adjust the energy distribution of the vibration waveform and output an inertial disturbance that matches the virtual physical law; When generating each frame of the virtual scene, the environmental evolution module maps the extreme points of the pressure gradient tensor to curvature singularities, driving the virtual scene's gravity field intensity distribution to adjust from 9.8 m / s² to 10.2 m / s², and increasing the material's viscoelastic coefficient by 15%. The virtual object's motion trajectory produces a nonlinear path deviation due to the correction of the geodesic equation, and the magnitude of the deviation is positively correlated with the real-time deformation of the dynamic actuation module. For example, when the deformation is 2 mm, the deviation is 5 cm. The cognitive coupling module monitors the Lyapunov exponent in real time. When the exponent exceeds a threshold of 0.5, the dynamic Bayesian network parameter recalibration process is triggered, pausing the prediction task and re-extracting the multidimensional feature vector of the latest physiological signal. The feature weights are adjusted in combination with the updated Lyapunov exponent, and the network model is retrained using the standard psychological state database to output a revised psychological state prediction value. The four-dimensional equilibrium state detection module continuously verifies whether the extreme point distribution of the heterogeneous sensing module and the deformation derivative of the dynamic actuation module satisfy the Hodge duality relationship, that is, the outer product of the two tensors is zero; at the same time, it checks whether the absolute value of the curvature distortion factor of the environmental evolution module is equal to the product of the psychological energy estimate and the preset proportional coefficient, the default value of the preset proportional coefficient is 0.5, and ensures that the ratio of the sampling rate of the heterogeneous sensing module 1000Hz to the rendering frame rate of the environmental evolution module 60Hz is a transcendental number (such as π); when all constraints are synchronously satisfied, the system outputs a valid control signal; if it is detected that the deformation derivative deviates from the threshold, the four-dimensional equilibrium state reconstruction process is immediately triggered, with the priority being: resetting the curvature distortion factor of the environmental evolution module to the initial proportional coefficient of 0.5; rolling back the deformation derivative of the dynamic actuation module to the last valid state, calling the historical deformation data buffer to restore the parameters; forcing the heterogeneous sensing module to reinitialize the data flow path and re-execute the data flow path reversal protocol; During the experiment, the predicted value of the subject's mental state is updated in real time through a dynamic Bayesian network. When the predicted value exceeds the safety threshold of 80, the priority arbitration mechanism of the dynamic actuation module forcibly suppresses the deformation rate to 0.2mm / s, and the acceleration of the electromagnetic damping motion base is reduced to 0.5g; at the same time, the actual acceleration data is fed back to the cognitive coupling module to calibrate the mental state prediction model to ensure that the virtual scene disturbance is synchronized with the subject's physiological response; at the end of the experiment, the system generates a comprehensive report, including the mental energy estimation value curve, the mental state migration path and the four-dimensional equilibrium state detection log, providing researchers with a basis for quantitative analysis.

[0025] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. A multifunctional seating system for psychological experiments and assessments with immersive experience, characterized by: It includes heterogeneous sensing module, dynamic actuation module, environmental evolution module, cognitive coupling module and four-dimensional equilibrium state detection; The heterogeneous sensing module includes a physiological signal acquisition unit and a metabolic parameter sensor array, which are used to synchronously capture heart rate variability, skin conductance response, surface electromyographic signals, and body surface temperature gradient data. It also uses a distributed pressure sensing array to capture the subject's body pressure distribution and micro-motion trajectory in real time, and converts the original mechanical signal into a pressure gradient tensor. When it is detected that the angle between the pressure gradient direction vector and the deformation direction vector of the dynamic actuation module reaches a preset threshold through vector dot product calculation, a data flow path reversal protocol is triggered to force the pressure gradient tensor to be transmitted to the environmental evolution module for preprocessing. The environmental evolution module maps the extreme points of the pressure gradient tensor to curvature singularities of the virtual space, and modifies the geodesic equation based on the deformation variable fed back by the dynamic actuation module to reconstruct the physical laws of the scene. Simultaneously, the inverse vector of the psychological state prediction value of the previous frame output by the cognitive coupling module is embedded as a curvature distortion factor. The inverse vector is generated by the cognitive coupling module through a gradient backprojection algorithm. Based on the modified geodesic equation and the curvature distortion factor, the second-order derivative of the Gaussian curvature of the virtual space is solved through a discretized Ricci flow algorithm to generate a curvature derivative instruction. The dynamic actuation module receives the curvature derivative instruction output by the environmental evolution module, dynamically adjusts the curvature of the supporting surface through the shape memory alloy actuator, and drives the electromagnetic damping motion base to generate inertial perturbations that match the physical laws of the virtual scene. Simultaneously, it receives the psychological state prediction value output by the cognitive coupling module, encodes it into the actuator deformation rate and the electromagnetic damping motion base acceleration, and combines it with the curvature derivative instruction to optimize the seat shape and movement pattern in real time, so that the inertial perturbation intensity generated by the seat is proportional to the subject's estimated psychological energy value. The cognitive coupling module generates a psychological energy estimation value by real-time analysis of the spectrum entropy of physiological signals and the mutation frequency of behavioral data, combined with the prediction results of the psychological state migration direction by the Lyapunov index. At the same time, a psychological state migration model is constructed through a dynamic Bayesian network to generate a psychological state prediction value. The psychological state prediction value is encoded as the motion parameter of the dynamic actuation module and the input parameter of the curvature distortion factor of the environmental evolution module. The curvature distortion factor of the environmental evolution module interacts with the temporal misalignment of the cognitive coupling module to form a recursive constraint among the four modules. The heterogeneous sensing module, dynamic actuation module, environmental evolution module, and cognitive coupling module achieve closed-loop coupling of data streams and control signals through four-dimensional equilibrium state detection. The four-dimensional equilibrium state detection includes a multi-dimensional constraint relationship between the extreme point distribution of the heterogeneous sensing module, the deformation derivative of the dynamic actuation module, the curvature distortion factor of the environmental evolution module, and the psychological energy estimation value of the cognitive coupling module. Only when all the constraint conditions are synchronously met is the system allowed to output a valid control signal.

2. The multifunctional seat system for psychological experiments and assessments with immersive experience function according to claim 1 is characterized by: The heterogeneous sensing module converts the discrete pressure values ​​captured by the distributed pressure sensing array into a pressure gradient tensor in a continuous space-time field, calculates the covariant derivatives of each sensing node on the Riemannian manifold to extract the distribution characteristics of saddle points and extreme points; when it is detected that the angle between the pressure gradient direction and the deformation direction vector of the dynamic actuation module reaches the preset threshold, the data flow path reversal protocol is triggered, so that the pressure gradient tensor is transmitted to the dynamic actuation module after being processed by the curvature singularity mapping of the environmental evolution module.

3. The multifunctional seat system for psychological experiments and assessments with immersive experience function according to claim 1 is characterized by: The dynamic actuation module receives the curvature derivative instruction output by the environmental evolution module, dynamically sets a deformation rate threshold of the shape memory alloy actuator based on the sign and amplitude of the derivative, and interprets the psychological state prediction value output by the cognitive coupling module into a deformation rate weighting factor and inertial disturbance waveform parameters; wherein the deformation rate threshold is adjusted in real time by the residual term of the geodesic equation solved by the environmental evolution module, the deformation rate weighting factor corrects the actual deformation of the actuator according to the nonlinear mapping rule of the psychological state prediction value, and sets a historical deformation data buffer to store the deformation parameters; The inertial disturbance waveform parameters generate a non-periodic vibration waveform based on the time series characteristics of the psychological state prediction value, and its amplitude is proportional to the psychological energy estimation value; at the same time, the dynamic actuation module coordinates the curvature derivative instruction and the input of the psychological state prediction value through a priority arbitration mechanism. When the psychological state prediction value exceeds the safety threshold, the deformation rate and acceleration are forcibly suppressed to ensure the safety of the subject.

4. The multifunctional seat system for psychological experiments and assessments with immersive experience function according to claim 3 is characterized by: The psychological state prediction value is decomposed into amplitude, frequency and phase components. The amplitude component is used to set the vibration intensity baseline, the frequency component controls the random distribution of the vibration interval, and the phase component is based on a chaotic phase superposition algorithm of the logistic map. At the same time, combined with the virtual scene inertia parameters extracted from the curvature derivative instruction, the energy distribution of the vibration waveform is dynamically adjusted to ensure that the inertial disturbance generated by the electromagnetic damping motion base is consistent with the physical laws of the virtual scene. The actual acceleration data is fed back to the cognitive coupling module to calibrate the psychological state prediction model.

5. The multifunctional seat system for psychological experiments and assessments with immersive experience function according to claim 1 is characterized by: When generating each frame of the virtual scene, the environmental evolution module maps the extreme points of the pressure gradient tensor output by the heterogeneous sensing module to curvature singularities of the virtual space through a discretized Ricci flow algorithm, and modifies the geometric parameter terms of the geodesic equation based on the real-time deformation variable, driving the evolution of the gravity field intensity distribution and the material viscoelastic coefficient of the virtual scene, so that the motion trajectory of the virtual object exhibits a nonlinear path deviation related to the deformation variable of the dynamic actuation module, and the deviation amplitude is positively correlated with the real-time deformation variable; the real-time deformation variable is generated by the deformation rate threshold of the shape memory alloy actuator; the extreme points are the concentrated or released areas of the subject's body pressure distribution of the pressure gradient tensor in the space-time field.

6. The multifunctional chair system for psychological experiments and assessments with immersive experience function according to claim 1 is characterized by: The cognitive coupling module extracts a multidimensional feature vector based on the spectrum entropy of the physiological signal collected by the heterogeneous sensing module and the mutation frequency of the behavioral data. Combined with the prediction result of the Lyapunov exponent on the direction of mental state migration, the cognitive coupling module constructs a mental state migration model through the dynamic Bayesian network and outputs a mental state prediction of the current frame synchronized with the rendering frame rate of the environmental evolution module. The psychological state prediction value is generated into an inverse vector through the gradient backprojection algorithm, and the inverse vector is used as the input parameter of the curvature distortion factor of the environmental evolution module; the migration direction of the subject's psychological state is predicted by calculating the Lyapunov exponent, and the sampling clock frequency of the heterogeneous sensing module is dynamically modified so that its ratio to the rendering frame rate of the environmental evolution module is an incommensurable transcendental number to avoid periodic interference; the dynamic Bayesian network performs real-time evaluation of the stability of the psychological state migration model based on the Lyapunov exponent, and when the Lyapunov exponent exceeds a preset threshold, the parameter recalibration process of the dynamic Bayesian network is triggered to ensure the convergence of the psychological state prediction value; the sampling clock frequency of the heterogeneous sensing module is the number of times physiological signals and mechanical data are collected per second; the rendering frame rate of the environmental evolution module is the number of frames updated per second for the virtual scene.

7. The multifunctional seat system for psychological experiments and assessments with immersive experience function according to claim 1 is characterized by: The four-dimensional equilibrium state detection includes: setting the extreme point distribution of the heterogeneous sensing module and the deformation derivative of the dynamic actuation module to satisfy the Hodge duality relationship; the absolute value of the curvature distortion factor of the environmental evolution module is equal to the product of the psychological energy estimate output by the cognitive coupling module and the preset proportional coefficient; and the ratio of the sampling rate of the heterogeneous sensing module to the rendering frame rate of the environmental evolution module is a transcendental number, allowing the system to output an effective response only when all conditions are met synchronously; the curvature distortion factor is generated by multiplying the inverse vector by the preset proportional coefficient, the psychological energy estimate is the weighted sum of the spectral entropy of the physiological signal and the mutation frequency of the behavioral data, and the weight is dynamically adjusted by the Lyapunov exponent; the Hodge duality relationship is defined as the constraint that the extreme point distribution tensor of the heterogeneous sensing module and the deformation derivative tensor of the dynamic actuation module satisfy the outer product constraint of zero.

8. The multifunctional chair system for psychological experiments and assessments with immersive experience function according to claim 1 is characterized by: The processing process when the four-dimensional equilibrium state is destroyed includes: when any constraint condition is detected to deviate from the threshold, the four-dimensional equilibrium state reconstruction process is immediately triggered, and the priority of the four-dimensional equilibrium state reconstruction process is: resetting the distortion factor generation rule of the environmental evolution module, including resetting the curvature distortion factor to the initial proportional coefficient; rolling back the deformation derivative of the dynamic actuation module to the last valid state, calling the historical deformation data buffer of the dynamic actuation module to restore the deformation parameters; forcing the heterogeneous sensing module to reinitialize the data flow path and re-execute the data flow path reversal protocol.

9. The multifunctional seat system for psychological experiments and assessments with immersive experience function according to claim 1 is characterized by: The method for generating the psychological state prediction value includes: performing wavelet denoising and normalization processing on the physiological signals collected by the heterogeneous sensing module; extracting time domain features and frequency domain features to construct a multidimensional feature vector; combining the prediction results of the psychological state migration direction by the Lyapunov index to correct the weight distribution of the multidimensional feature vector; inputting the feature vector into a dynamic Bayesian network pre-trained based on a standard psychological state database to output a psychological state migration probability distribution; generating a psychological state prediction value of the current frame according to the maximum likelihood estimate of the psychological state migration probability distribution; the maximum likelihood estimate is the numerical value corresponding to the psychological state category with the highest probability.