Attitude sensing method and device based on flexible wearable robot

Abstract

The invention relates to a posture sensing method and device based on a flexible wearable robot, and belongs to the technical field of rehabilitation robot application, and the method comprises the steps: synchronously collecting standard motion data and real-time motion data of a user; performing coarse attitude calculation through a complementary filtering algorithm to obtain initial six-degree-of-freedom attitude data; calibrating the initial six-degree-of-freedom attitude data based on the deformation resistance change data, the real-time environment temperature data and a pre-constructed attitude deformation calibration model, completing attitude drift correction and motion artifact elimination, and obtaining calibrated six-degree-of-freedom attitude data; based on the calibrated six-degree-of-freedom attitude data, extracting user exclusive action characteristics and temperature correlation characteristics, and dynamically correcting core parameters of the attitude deformation calibration model according to the user exclusive action characteristics and the temperature correlation characteristics; and inputting the real-time action data into the corrected attitude deformation calibration model, and outputting real-time attitude data. The method has the effect of improving the attitude sensing precision and real-time performance.

Description

Technical Field

[0001] This application relates to the technical field of rehabilitation robot applications, and in particular to a posture perception method and device based on a flexible wearable robot. Background Technology

[0002] In the field of flexible wearable robots, high-precision posture perception is a core prerequisite for achieving accurate human body capture, dexterous operation assistance, and rehabilitation training guidance. Its performance directly determines the interactive reliability and application experience of wearable robots. In the existing posture perception of flexible wearable robots, the deformation of the flexible substrate is generally regarded as a negative interference, and adversarial design is adopted to try to suppress or eliminate the influence of deformation. However, this leads to a trade-off between wearing comfort and posture accuracy.

[0003] Existing solutions address motion artifacts and attitude drift caused by deformation. On one hand, they reduce deformation by strengthening the rigidity of the flexible substrate, which improves attitude perception accuracy but compromises the fit and comfort of the device. On the other hand, they employ a multi-hardware overlay scheme and use complex fusion algorithms to counteract deformation interference. However, these fusion algorithms consume a lot of computing power and cannot meet the real-time requirements of precise operations.

[0004] Therefore, there is an urgent need for a posture sensing solution that can fully utilize the deformation characteristics of flexible substrates to achieve high precision and real-time performance. Summary of the Invention

[0005] To improve the accuracy and real-time performance of posture perception, this application provides a posture perception method and device based on a flexible wearable robot.

[0006] Firstly, this application provides a posture perception method based on a flexible wearable robot, employing the following technical solution: The system simultaneously collects the user's standard motion data and real-time motion data. The standard motion data includes the overall motion acceleration data of the wearable part, angular velocity data, deformation resistance change data of the flexible substrate, and real-time ambient temperature data. Based on the overall motion acceleration data and the angular velocity data, coarse attitude calculation is performed using a complementary filtering algorithm to obtain initial six-degree-of-freedom attitude data. The initial six-degree-of-freedom attitude data is calibrated based on the deformation resistance change data, the real-time ambient temperature data, and the pre-constructed attitude deformation calibration model to complete attitude drift correction and motion artifact removal, thereby obtaining calibrated six-degree-of-freedom attitude data. Based on the calibrated six-degree-of-freedom attitude data, user-specific motion features and temperature-related features are extracted, and the core parameters of the attitude deformation calibration model are dynamically corrected according to the user-specific motion features and the temperature-related features. The real-time motion data is input into the corrected attitude deformation calibration model, and real-time attitude data is output.

[0007] By adopting the above technical solution, through the whole process design of multi-dimensional data synchronous acquisition, complementary filtering coarse calculation, deformation temperature joint calibration and model dynamic correction, the deformation of flexible substrate is transformed from an interference source into a calibration basis. Combined with temperature compensation, the attitude drift and motion artifact errors are greatly reduced, the attitude angle perception accuracy is improved, and the model parameters are dynamically corrected based on user-specific features to achieve personalized adaptation.

[0008] Furthermore, the calibration of the initial six-degree-of-freedom attitude data based on the deformation resistance change data, the real-time ambient temperature data, and the pre-constructed attitude deformation calibration model, to complete attitude drift correction and motion artifact removal, includes: The pre-built attitude deformation calibration model expression is: ; in, The calibrated attitude angles The angular components are the initial six-DOF attitude data, and k is the coupling coefficient. The rate of change of resistance is the ratio of the resistance change measured by the strain-sensitive grid to the initial resistance. For temperature compensation, For real-time ambient temperature, For calibration reference temperature; Substitute the angle components, deformation resistance change data, and real-time ambient temperature data from the initial six-degree-of-freedom attitude data into the model to obtain the output calibrated six-degree-of-freedom attitude data. If the amplitude of attitude angle change in the initial six-degree-of-freedom attitude data or the calibrated six-degree-of-freedom attitude data is greater than the artifact detection threshold δ, it is determined to be a motion artifact and the initial six-degree-of-freedom attitude data or the calibrated six-degree-of-freedom attitude data is removed.

[0009] By adopting the above technical solution, using the deformation temperature joint calibration model, combined with artifact threshold judgment, applying flexible deformation data to compensate for initial attitude deviation, offsetting environmental interference through temperature terms, and eliminating abrupt artifact data, the accuracy of attitude angle measurement is significantly improved, ensuring the data is true and reliable.

[0010] Furthermore, the user-specific motion features include joint range of motion features, deformation rate features, posture deformation correlation features, and motion stability features. The core parameters of the posture deformation calibration model, dynamically adjusted based on the user-specific motion features and the temperature correlation features, include: The maximum bending angle of the user's joint is obtained based on the joint range of motion characteristics, the average deformation rate of the user's joint is obtained based on the deformation rate characteristics, the posture deformation correlation coefficient is obtained based on the posture deformation correlation characteristics, the posture angle standard deviation of repeated actions is obtained based on the action stability characteristics, and the real-time ambient temperature and temperature change rate are obtained based on the temperature correlation characteristics. Based on the user-specific motion characteristics, the coupling coefficient k, the temperature compensation term ε, and the artifact detection threshold δ of the posture deformation calibration model are adjusted by a quantization correction formula. The correction formula for the coupling coefficient k is: ; in, This is the standard factory coupling coefficient. The maximum bending angle of the user's joint. This represents the maximum bending angle of a standard joint. The correlation coefficient for user posture deformation. The standard correlation coefficient is used. To calibrate the reference temperature, For real-time ambient temperature, , These are the weighting coefficients for action features. These are the temperature characteristic weighting coefficients; The correction formula for the temperature compensation term ε is: ; in, This is a general temperature compensation value. The average deformation rate of the user's joint. The standard average deformation rate is used. For the rate of temperature change, For the general standard rate of temperature change, This is the deformation rate weighting coefficient. These are the temperature characteristic weighting coefficients; The correction formula for the artifact detection threshold δ is: ; in, This is a general artifact detection threshold. The standard deviation of the posture angles of repeated user actions. The standard deviation of the universal standard motion. These are the weighting coefficients.

[0011] By adopting the above technical solution, based on the user's joint movement, deformation rate and other unique features and temperature-related features, the three core parameters of coupling coefficient, temperature compensation term and artifact threshold are dynamically corrected to achieve personalized model adaptation, effectively offset individual movement differences and environmental temperature interference, and significantly improve the accuracy and stability of posture calibration.

[0012] Furthermore, the overall motion acceleration data and the angular velocity data are both acquired by an IMU chip, and the deformation resistance change data are acquired by a strain-sensitive grid. The method also includes: The posture data is matched with the action trend mapping library to obtain the current posture change trend and predict the user's action complexity level within a preset time. The action complexity level is determined by a weighted average of the posture angle change rate and the number of joints involved. Based on the predicted action complexity level, the sampling scheme is adjusted in reverse: If the complexity level of the action is low, then reduce the sampling accuracy of the IMU chip, decrease the amplification factor of the strain-sensitive gate signal, and disable redundant operations of the IMU chip; If the complexity level of the action is medium, then the sampling accuracy of the IMU chip and the amplification factor of the strain-sensitive gate are maintained, and the basic operation of the IMU chip is activated. If the complexity level of the action is high, the sampling accuracy of the IMU chip is increased, the amplification factor of the strain-sensitive gate signal is increased, all computing units of the IMU chip are activated, and the coupling calibration model parameters are preloaded.

[0013] By adopting the above technical solution, the complexity level is predicted through action trend matching, and the sampling scheme is dynamically adjusted in reverse: low-level actions have reduced accuracy and redundancy disabled, mid-level actions maintain basic configuration, and high-level actions have increased accuracy and run at full load. This reduces the power consumption of simple actions while ensuring the real-time perception and accuracy of complex actions, achieving an optimal balance between computing power and performance.

[0014] Furthermore, the method also includes: The user's personalized feature data is collected in advance, and a user fit coefficient K is constructed based on the personalized feature data; wherein, the user fit coefficient K is dynamically updated as the user's recovery process progresses; Real-time posture data of the target rehabilitation area during rehabilitation movements is obtained and compared with standard rehabilitation movements to obtain the initial movement deviation value. The initial correction intensity is calculated based on the user adaptation coefficient K. The initial correction intensity is taken as the current correction intensity; wherein, Basic strength; The flexible actuator module outputs correction force according to the current correction intensity; Collect posture feedback data, calculate the residual deviation between the posture feedback data and standard rehabilitation movements, and determine the intensity adjustment step size based on the correlation between the residual deviation and the user fit coefficient K. The current correction intensity and the intensity adjustment step size are set. The updated correction strength is obtained by adding them together, and the updated correction strength is used as the new current correction strength. The repetitive execution-driven flexible execution module outputs a corrective force according to the current corrective intensity; it collects posture feedback data, calculates the residual deviation between the posture feedback data and the standard rehabilitation movement, and determines the intensity adjustment step size based on the correlation between the residual deviation and the user fit coefficient K. Adjust the step size based on the current correction intensity and the intensity. The updated correction intensity is obtained, and the updated correction intensity is used as the new current correction intensity step until the remaining deviation is less than or equal to the preset threshold, and the final correction intensity is determined.

[0015] By adopting the above technical solution, an adaptation coefficient is constructed based on the user's personalized characteristics. The intensity of the correction is dynamically calculated and iteratively optimized in combination with the movement deviation. The correction is adjusted in real time to adapt to the rehabilitation process. The flexible execution module outputs precise correction force, gradually reducing the movement deviation to the target threshold, thereby realizing personalized and refined rehabilitation training and improving the safety and effectiveness of training.

[0016] Furthermore, the intensity adjustment step size is determined based on the correlation between the residual deviation and the user adaptation coefficient K. Adjust the step size based on the current correction intensity and the intensity. The updated correction strength is obtained, including: Calculate the product of the remaining deviation and the user adaptation coefficient K to obtain the correlation relationship; If the association relationship is greater than the association threshold, then according to Adjust the step size, and compare the current correction intensity with... Add them together to get the updated correction strength; If the association relationship is less than or equal to the association threshold, then according to... Adjust the step size, and compare the current correction intensity with... Subtracting the two gives the updated correction strength; Where i represents the number of adjustment steps, i∈[1,n]. When i=1, ... Use the base step size.

[0017] By adopting the above technical solution, the correlation is determined by the product of the residual deviation and the user adaptation coefficient. Combined with the dynamic adjustment of the correction step length during the rehabilitation cycle, if the deviation is large, the correction intensity is increased proportionally to the original step length; if the deviation is small, the step length is halved to reduce the intensity. This achieves precise iterative optimization of the correction force, ensuring the safety and effectiveness of rehabilitation training.

[0018] Furthermore, the method also includes: The personalized feature data includes the maximum voluntary contractile force of the user's target area. Maximum range of motion of the joint The process of constructing the user adaptation coefficient K based on the personalized feature data includes: ,in, This represents the average maximum voluntary contractile force of healthy individuals in the corresponding age group. This represents the average maximum fluctuating range of motion of joints in healthy individuals within the corresponding age group. Real-time compensation for the user adaptation coefficient K includes: If after two consecutive rounds of adjustment, the absolute value of the rate of change of the residual deviation If the preset rate is reached, the user adaptation coefficient K will be compensated and corrected. If the rate of change of residual deviation If it is negative, then ; If the rate of change of residual deviation If it is positive, then ;in, This is the remaining deviation; The intensity adjustment step size is recalculated based on the corrected user adaptation coefficient. .

[0019] By adopting the above technical solutions, personalized adaptation coefficients are constructed based on muscle strength and joint range of motion, prioritizing the matching of differences in joint mobility; combined with the dynamic compensation correction coefficient of the residual deviation change rate, if the deviation decreases too quickly, the coefficient is reduced to stabilize the correction, and if the deviation increases too quickly, the coefficient is increased to strengthen the correction, thereby achieving precise and personalized adaptation of rehabilitation correction force and improving the safety and effectiveness of training.

[0020] Secondly, this application provides a posture sensing device based on a flexible wearable robot, employing the following technical solution: The data acquisition module is used to simultaneously collect user standard motion data and real-time motion data. The standard motion data includes the overall motion acceleration data of the wearable part, angular velocity data, deformation resistance change data of the flexible substrate, and real-time ambient temperature data. The initial six-degree-of-freedom attitude data calculation module is used to perform coarse attitude calculation based on the overall motion acceleration data and the angular velocity data, and obtain the initial six-degree-of-freedom attitude data by using a complementary filtering algorithm. The attitude correction and motion artifact removal module is used to calibrate the initial six-degree-of-freedom attitude data based on the deformation resistance change data, the real-time ambient temperature data, and the pre-built attitude deformation calibration model, to complete attitude drift correction and motion artifact removal, and obtain calibrated six-degree-of-freedom attitude data. The model calibration module is used to extract user-specific motion features and temperature-related features based on the calibrated six-degree-of-freedom attitude data, and dynamically correct the core parameters of the attitude deformation calibration model according to the user-specific motion features and the temperature-related features. The attitude data output module is used to input the real-time motion data into the corrected attitude deformation calibration model and output real-time attitude data.

[0021] By adopting the above technical solution, through the whole process design of multi-dimensional data synchronous acquisition, complementary filtering coarse calculation, deformation temperature joint calibration and model dynamic correction, the deformation of flexible substrate is transformed from an interference source into a calibration basis. Combined with temperature compensation, the attitude drift and motion artifact errors are greatly reduced, the attitude angle perception accuracy is improved, and the model parameters are dynamically corrected based on user-specific features to achieve personalized adaptation.

[0022] Thirdly, this application provides a flexible wearable robot, which adopts the following technical solution: The flexible composite sensing layer comprises, from top to bottom, a polyimide substrate, a silver nanowire conductive network, and a PDMS encapsulation layer; the flexible composite sensing layer integrates a triaxial MEMS IMU chip, a strain-sensitive gate, and a temperature compensation unit. The data processing module includes a SoC chip integrating an ARM Cortex M7 core and an FPGA coprocessor; the flexible composite sensing layer is electrically connected to the data processing module. A flexible execution module, which communicates with the data processing module via a wireless communication module; Memory; At least one computer program, wherein the at least one computer program is stored in the memory and configured to be executed by the data processing module, the at least one computer program being configured to: perform the posture perception method based on a flexible wearable robot as described in any one of the first aspects.

[0023] By adopting the above technical solution, this flexible wearable robot uses a polyimide-silver nanowire-PDMS flexible composite sensing layer, which conforms to human skin and has strong deformation adaptability, and can simultaneously collect multi-dimensional motion and environmental data. The heterogeneous data processing module of ARM Cortex-M7 and FPGA enables high-speed parallel computation of posture calculation, calibration, and correction, balancing high precision and low latency. The flexible execution module, combined with a personalized posture perception method, outputs precise flexible correction force, adapting to the needs of rehabilitation training scenarios. The overall device is comfortable to wear, responds quickly, and controls precisely, effectively solving the problems of poor comfort and unstable accuracy of traditional rigid wearable devices.

[0024] Fourthly, this application provides a computer-readable storage medium, which adopts the following technical solution: A computer-readable storage medium storing a computer program capable of being loaded by a processor and executed as described in any one of the first aspects, a posture perception method based on a flexible wearable robot.

[0025] By adopting the above technical solution, the processor loads a computer program in a computer-readable storage medium. Through the full-process design of multi-dimensional data synchronous acquisition, complementary filtering coarse calculation, deformation temperature joint calibration and model dynamic correction, the deformation of the flexible substrate is transformed from an interference source into a calibration basis. Combined with temperature compensation, the attitude drift and motion artifact errors are greatly reduced, the attitude angle perception accuracy is improved, and the model parameters are dynamically corrected based on user-specific features to achieve personalized adaptation.

[0026] In summary, this application includes at least one of the following beneficial technical effects: 1. Through the full-process design of multi-dimensional data synchronous acquisition, complementary filtering coarse calculation, deformation temperature joint calibration and model dynamic correction, the deformation of flexible substrate is transformed from an interference source into a calibration basis. Combined with temperature compensation, the attitude drift and motion artifact errors are greatly reduced, the attitude angle perception accuracy is improved, and the model parameters are dynamically corrected based on user-specific features to achieve personalized adaptation. 2. Low-level actions reduce accuracy and eliminate redundancy, mid-level actions maintain basic configuration, and high-level actions improve accuracy and run at full load; this reduces the power consumption of simple actions while ensuring the real-time perception and accuracy of complex actions, achieving the optimal balance between computing power and performance. 3. Optimize the correction intensity based on the user's personalized characteristics and adjust it in real time to adapt to the rehabilitation process, so as to realize personalized and refined rehabilitation training and improve the safety and effectiveness of training. Attached Figure Description

[0027] Figure 1 This is a flowchart illustrating the posture perception method based on a flexible wearable robot in an embodiment of this application.

[0028] Figure 2 This is a structural block diagram of the posture sensing device based on a flexible wearable robot in the embodiments of this application. Detailed Implementation

[0029] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0030] Furthermore, the term "and / or" in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. Additionally, the character " / " in this article, unless otherwise specified, generally indicates that the preceding and following related objects have an "or" relationship.

[0031] This application discloses a flexible wearable robot, including a flexible composite sensing layer, a data processing module, a flexible execution module, and a memory.

[0032] The flexible composite sensing layer consists of a polyimide substrate, a silver nanowire conductive network, and a PDMS encapsulation layer from top to bottom. It integrates a triaxial MEMS IMU chip, strain-sensitive gates, and a temperature compensation unit, all electrically connected to the data processing module via flexible circuitry. The flexible circuitry uses printed circuitry on a polyimide substrate, allowing it to bend and stretch with the user's movements. The triaxial MEMS IMU chip is positioned at the posture reference point of the wearable device, i.e., the location with minimal motion deformation. Strain-sensitive gate arrays are arranged at both the proximal and distal joints along the direction of motion, with two sets of orthogonal strain-sensitive gates corresponding to each joint, arranged along the bending direction and perpendicular to the bending direction respectively, ensuring comprehensive capture of joint deformation. The temperature compensation unit is a temperature thermistor, which is attached to the flexible base area adjacent to the triaxial MEMS IMU chip.

[0033] The data processing module uses a SoC chip integrating an ARM Cortex-M7 core and an FPGA coprocessor. The ARM Cortex-M7 core is responsible for algorithm logic control and parameter correction, while the FPGA coprocessor is responsible for high-speed data preprocessing and calibration calculations. The two communicate via an internal bus. The data processing module also integrates a power management module to adapt to the power supply of the wearable robot.

[0034] The flexible actuator module uses a pneumatic flexible airbag, which is deployed on the outside of the joint of the moving part. It is connected to a micro air pump through an air tube. The micro air pump is controlled by the SoC chip through a wireless communication module. Based on the correction intensity parameters output by the attitude perception method, it precisely adjusts the pressure inside the airbag to achieve flexible correction.

[0035] The memory stores a computer program configured to be executed by the data processing module, which automatically completes the attitude perception and correction control process after startup.

[0036] This application discloses a posture perception method based on a flexible wearable robot. (Refer to...) Figure 1 This is performed by a flexible wearable robot. It includes (steps S101 to S105): Step S101: Synchronously collect user standard motion data and real-time motion data. The standard motion data includes the overall motion acceleration data of the worn parts, angular velocity data, deformation resistance change data of the flexible substrate, and real-time ambient temperature data.

[0037] Specifically, the flexible wearable robot collects data through a flexible composite sensing layer. The three-axis MEMS IMU chip collects overall motion acceleration and angular velocity data. When the user's joints bend, the flexible substrate deforms with the joint, and the resistance of the silver nano-conductive network changes. The resistance is converted into a voltage signal through a signal conditioning circuit, and the strain-sensitive gate collects the deformation resistance change data. When the user wears the robot, the thermistor records the temperature of the wearing environment and the skin contact in real time to obtain real-time ambient temperature data.

[0038] The standard movement data is collected when the user completes the standard movements of rehabilitation training. It is obtained by taking the average value after each standard movement is collected multiple times.

[0039] Step S102: Based on the overall motion acceleration data and angular velocity data, coarse attitude calculation is performed using a complementary filtering algorithm to obtain the initial six-degree-of-freedom attitude data.

[0040] Specifically, the complementary filtering algorithm has a small computational load and can be executed quickly on low-computing-power hardware.

[0041] The collected overall motion acceleration and angular velocity data are input into the complementary filtering algorithm: ; in, These are the attitude angles after coarse calculation, including roll, pitch, and yaw angles, which are the angular components of the initial six-DOF attitude data. These are complementary coefficients, which can be calibrated based on motion characteristics. The attitude angle is obtained by integrating the gyroscope. For angular velocity data, The sampling time interval, The attitude angle is calculated based on the overall motion acceleration data.

[0042] Step S103: Based on the deformation resistance change data, real-time ambient temperature data, and the pre-built attitude deformation calibration model, calibrate the initial six-degree-of-freedom attitude data to complete attitude drift correction and motion artifact removal, including: Specifically, the pre-constructed posture deformation calibration model reflects a fixed mapping relationship between the deformation resistance change of the flexible substrate and the joint posture angle, and this mapping relationship is calibrated through standard motion data.

[0043] The model expression is: ; in, The calibrated attitude angles The angular components are the initial six-DOF attitude data, and k is the coupling coefficient. The rate of change of resistance is the ratio of the resistance change measured by the strain-sensitive grid to the initial resistance. For temperature compensation, For real-time ambient temperature, This is for calibrating the reference temperature.

[0044] For deformation compensation, the deformation characteristics of the flexible substrate are used to assist in calibration. The deformation resistance is the rate of change; the larger the joint bending angle, the more pronounced the deformation. The larger the value of , the greater the correlation coefficient between deformation and attitude. Therefore, comparing K with... Multiplication is performed to deduce the attitude angle components missed in the coarse calculation using deformation data, and these components are positively superimposed onto the initial values ​​to cancel out motion artifacts caused by deformation.

[0045] This is a temperature correction term to counteract the interference of temperature on the sensing element. The deviation between real-time temperature and the reference. Quantizing the effect of temperature drift on attitude angle. That is, when the temperature rises, If the value is positive, the additional attitude angle error caused by temperature deviation is subtracted. When hour, If the value is negative, it will positively compensate for the attitude angle deviation caused by the decrease in temperature.

[0046] Therefore, the attitude deformation calibration model corrects the coarsely calculated attitude angles to precise attitude angles, utilizing deformation and addressing temperature offset issues. During calibration, the angular components, deformation resistance changes, and real-time ambient temperature data from the initial six-degree-of-freedom attitude data are substituted into the model to calibrate the initial left-degree-of-freedom attitude data.

[0047] However, motion artifacts may be caused by device sliding or involuntary muscle tremors. Therefore, an artifact detection threshold is set. When the change in attitude angle in the initial six-degree-of-freedom attitude data or the calibrated six-degree-of-freedom attitude data exceeds the artifact detection threshold, it is identified as a motion artifact and the initial six-degree-of-freedom attitude data or the calibrated six-degree-of-freedom attitude data is removed.

[0048] Step S104: Based on the calibrated six-DOF attitude data, extract user-specific motion features and temperature-related features, and dynamically correct the core parameters of the attitude deformation calibration model according to the user-specific motion features and temperature-related features. User-specific motion features include joint range of motion features, deformation rate features, attitude deformation correlation features, and motion stability features. This includes steps S1041 to S1042.

[0049] Step S1041: Obtain the maximum bending angle of the user joint based on the joint range of motion characteristics, obtain the average deformation rate of the user joint based on the deformation rate characteristics, obtain the posture deformation correlation coefficient based on the posture deformation correlation characteristics, obtain the posture angle standard deviation of repeated actions based on the action stability characteristics, and obtain the real-time ambient temperature and temperature change rate based on the temperature correlation characteristics.

[0050] Specifically, the joint range of motion feature is the user's joint posture angle data, from which the maximum bending angle of the user's joint can be obtained; the deformation rate feature is the rate at which the user completes each action, which is calculated by the time of deformation resistance change data, and thus the average deformation rate of the user's joint can be calculated; the posture deformation correlation feature is the relationship between posture angle change and corresponding deformation resistance change, and the user posture deformation correlation coefficient is obtained by calculating the Pearson correlation coefficient between posture angle change and deformation resistance change; the action stability feature is the posture angle of the user each time a standard action is repeated, and the standard deviation of multiple posture angles can be calculated; the temperature correlation feature is the temperature data during the user's execution of the standard action, so the real-time ambient temperature and temperature change rate can be obtained.

[0051] Step S1042: Based on user-specific motion features, adjust the coupling coefficient k, the correction coefficient α of the temperature compensation term ε, and the artifact detection threshold δ of the attitude deformation calibration model through a quantization correction formula.

[0052] (1) The correction formula for the coupling coefficient k is: ; in, This is the standard factory coupling coefficient. The maximum bending angle of the user's joint. This represents the maximum bending angle of a standard joint. The correlation coefficient for user posture deformation. The standard correlation coefficient is used. To calibrate the reference temperature, For real-time ambient temperature, , These are the weighting coefficients for action features. This represents the temperature characteristic weighting coefficient.

[0053] Specifically, the coupling coefficient k is the correlation coefficient between the deformation resistance change of the flexible substrate and the joint attitude angle; essentially, it establishes the mapping ratio between the deformation signal and the attitude signal. Because different users have differences in joint range of motion, movement continuity, and skin fit, and because temperature affects the deformation and resistance response characteristics of flexible materials, a fixed factory-set coupling coefficient cannot be guaranteed. It cannot be accurately adapted to all scenarios.

[0054] In the correction formula for the coupling coefficient k, This is a joint range of motion correction item. Different users have different joint flexibility, resulting in different posture angles under the same deformation. If < ,but For negative numbers, through weaken To avoid overestimating the user's posture angle and ensure that it matches the user's actual joint mobility.

[0055] This is a correction term related to attitude deformation. Reflecting the continuity of deformation and posture in user actions. The larger the value, the stronger the connection between the two. If the user > ,but If it is a positive number, then through Enhance This increases the contribution of deformation signals to attitude calibration, improving calibration accuracy; conversely, if user actions are inconsistent, < Then weaken This reduces interference from invalid deformation signals.

[0056] This is a temperature correction term. The elastic modulus of flexible materials changes with temperature, and the change in resistance at the same bending angle will shift with temperature. For example, as the temperature rises, the material softens, and the change in resistance will be greater at the same bending angle. Therefore, adjustments can be made based on the deviation between the real-time temperature and the reference temperature. This compensates for the effect of temperature on the deformation-resistance response characteristics and reduces calibration errors caused by temperature deviation.

[0057] You can choose 0.4. You can choose 0.3. You can choose 0.3.

[0058] (2) The correction formula for the temperature compensation term ε is: ; in, This is a general temperature compensation value. The average deformation rate of the user's joint. The standard average deformation rate is used. For the rate of temperature change, For the general standard rate of temperature change, This is the deformation rate weighting coefficient. This represents the temperature characteristic weighting coefficient.

[0059] Specifically, the temperature compensation term ε is the attitude angle correction amount corresponding to a unit temperature change, and its function is to offset the influence of temperature on sensing components such as the IMU chip and strain-sensitive gate. Since different users have different action rates, and the action rate affects the dynamic response characteristics of the sensing components, and the real-time temperature change rate varies, a fixed term cannot accurately compensate for temperature errors in all scenarios. Therefore, it is necessary to dynamically fine-tune the temperature compensation term based on the user's action rate characteristics and temperature change rate characteristics.

[0060] This is a correction term for the action rate. The faster the action rate, the more lagging the dynamic response of the strain-sensitive gate, and the greater the influence of temperature. If > ,but If it is a positive number, then... Enhance This means the temperature compensation is more sensitive, effectively offsetting temperature errors caused by excessively rapid movements. If the user's movements are slow, the sensitivity is reduced. Reduce overcompensation.

[0061] This is a correction term for the rate of temperature change. The faster the rate of temperature change, the more drastic the changes in the performance parameters of the sensing element, and the faster the corresponding attitude angle temperature error accumulates. The value of ε is dynamically adjusted by comparing the ratio of the real-time temperature change rate to the standard value. > At that time, through Significantly enhanced It quickly compensates for errors caused by rapid temperature changes; when the temperature stabilizes, it weakens the error. The magnitude of the enhancement should be adjusted to avoid unnecessary over-correction.

[0062] Since temperature compensation is mainly used to address dynamic changes in ambient temperature. 0.2 can be selected. A value of 0.8 can be selected to make the influence of the temperature change rate characteristic much greater than that of the action rate characteristic.

[0063] (3) The correction formula for the artifact detection threshold δ is: ; in, This is a general artifact detection prevalue. The standard deviation of the posture angles of repeated user actions. The standard deviation of the universal standard motion. These are the weighting coefficients.

[0064] Specifically, the artifact detection threshold δ is the critical value for detecting abrupt changes in posture angles. It distinguishes between posture angle changes resulting from valid movements and abrupt changes in posture angles caused by motion artifacts. For example, abrupt changes in posture angles caused by non-target movements such as slight slippage of wearable devices or involuntary muscle tremors are considered motion artifacts and must be removed.

[0065] Because different users have different levels of action stability, a fixed... It cannot accurately adapt to all users. If it's too small, it will misinterpret normal fluctuations in the user's movements as artifacts; if If the value is too large, it will miss real motion artifacts, both of which will affect the accuracy of posture perception. Therefore, it is necessary to use user motion stability features to... Dynamic fine-tuning is performed to ensure the accuracy of artifact detection.

[0066] It reflects the degree of fluctuation in user actions. As a correction measure, adjustments are made dynamically by comparing the stability of user actions with general standards. The size. If > If the movements are unstable, then If it is a positive number, then... Increase To avoid misinterpreting normal user movement fluctuations as artifacts, if If the action is stable, the correction term is negative, decreasing the value. This improves the sensitivity of artifact detection.

[0067] Step S105: Input the real-time motion data into the corrected attitude deformation calibration model and output the real-time attitude data.

[0068] Specifically, the corrected attitude deformation calibration model is a model that fits the user's situation. After real-time motion data is input into the model, it can output accurate real-time attitude data.

[0069] Furthermore, the method also includes steps S11 to S12: Step S11: Match the posture data with the motion trend mapping library to obtain the current posture change trend and predict the user's motion complexity level within a preset time period. The motion complexity level is determined by a weighted average of the posture angle change rate and the number of joints involved.

[0070] Specifically, the motion trend mapping library pre-stores the posture change trends of various motions. For example, in finger activities, the posture angle of the finger-attacking motion changes rapidly and involves a large number of joints, making it a complex motion; while the posture angle of the extension motion changes slowly and involves fewer joints, making it a simple motion.

[0071] Specifically, the complexity level is determined by weighting the rate of change of attitude angle and the number of joints involved. The complexity level ranges from 0 to 10 and is divided into three motion complexity levels.

[0072] Step S12: Adjust the sampling scheme in reverse based on the predicted action complexity level: If the complexity level of the action is low, reduce the sampling accuracy of the IMU chip, decrease the amplification factor of the strain-sensitive gate signal, and disable redundant chip operations. If the complexity level is medium, then maintain the sampling accuracy of the IMU chip and the amplification factor of the strain-sensitive gate, and activate the chip's basic operations. If the complexity level of the operation is high, the sampling accuracy of the IMU chip is increased, the amplification factor of the strain-sensitive gate signal is increased, all computing units of the chip are activated, and the parameters of the coupling calibration model are preloaded.

[0073] By adjusting the sampling scheme dynamically, a balance can be struck between real-time performance and power consumption.

[0074] In another possible implementation, existing rehabilitation training often fails to adapt the correction intensity to individual user differences and rehabilitation progress, leading to poor training results or injury risks. This application addresses this by constructing a user fit coefficient to achieve personalized dynamic adjustment of the correction intensity. The method also includes steps S21 to S24: Step S21: Collect personalized feature data of users in advance, and construct user fit coefficient K based on personalized feature data; wherein, user fit coefficient K is dynamically updated as the user's recovery process progresses.

[0075] Personalized feature data includes the maximum voluntary contractile force at the user's target site. Maximum range of motion of the joint The user adaptation factor K includes: .

[0076] in, This represents the average maximum voluntary contractile force of healthy individuals in the corresponding age group. This represents the average maximum fluctuating range of motion of joints in healthy individuals within the corresponding age group. The user's maximum voluntary contractile force was calculated compared to that of healthy individuals of the same age. The maximum fluctuating joint range of motion of the user compared to healthy individuals of the same age was calculated. Joint range of motion directly determines whether the user can complete the standard movement and is a core factor affecting the correction effect. For example, users with limited joint mobility require less correction force to avoid overstretching. Muscle strength mainly affects the stability of movement completion and has a relatively minor impact on the correction intensity. Through weight allocation, K can more accurately match the user's core ability shortcomings in completing the movement.

[0077] If a user's muscle strength and joint range of motion are lower than those of healthy individuals, using the standard correction force for healthy individuals may lead to muscle strain or training phobia because the user cannot tolerate it. However, through K's quantitative adaptation, the correction intensity can be adjusted to a safe range that the user can tolerate, thereby improving the safety and effectiveness of rehabilitation training.

[0078] Step S22: Obtain real-time posture data of the target rehabilitation area during rehabilitation movements and compare it with standard rehabilitation movements to obtain the initial movement deviation value. The initial correction intensity is calculated by combining the user adaptation coefficient K. The initial correction strength is used as the current correction strength, where, Based on the strength.

[0079] Specifically, by multiplying the three values, the three key factors of individual differences, standard baseline strength, and actual deviation can be integrated to obtain the force that the user needs to bring back the deviating action within an acceptable range.

[0080] Step S23: Drive the flexible execution module to output the correction force according to the current correction intensity.

[0081] Step S24: Collect posture feedback data, calculate the residual deviation between the posture feedback data and the standard rehabilitation movement, and determine the intensity adjustment step size based on the correlation between the residual deviation and the user fit coefficient K. The current correction intensity and intensity adjustment step size are adjusted. The sums are used to obtain the updated correction intensity, which is then used as the new current correction intensity. Steps S23 and S24 are repeated until the remaining deviation is less than or equal to the preset threshold, at which point the final correction intensity is determined. At this intensity, the user can stably complete the standard gripping action without excessive or insufficient force.

[0082] Specifically, after the user performs the same movement 10 times consecutively, posture feedback data is collected and processed to obtain the average deviation between the posture feedback data and the standard rehabilitation movement in the last few movements as the residual deviation.

[0083] The intensity adjustment step size is determined based on the correlation between the residual deviation and the user adaptation coefficient K. Adjust the step size according to the intensity. The updated correction strength is obtained, including steps Sa to Sc: Step Sa: Calculate the product of the residual deviation and the user adaptation coefficient K to obtain the correlation.

[0084] Step Sb: If the association relationship is greater than the association threshold, then proceed according to... Adjust the step size, and compare the current correction intensity with... Add them together to get the updated correction strength.

[0085] Step Sc: If the association relationship is less than or equal to the association threshold, then proceed according to... Adjust the step size, and compare the current correction intensity with... Subtracting the two values ​​yields the updated correction strength. Here, i represents the current number of rehabilitation cycles, i∈[1,n]. When i=1, Use the base step size.

[0086] Specifically, the correlation threshold can be determined experimentally. The product of the remaining deviation and the user adaptation coefficient K reflects whether the correction intensity meets the requirements. If the correlation is greater than the correlation threshold, the current remaining deviation is relatively large, or the user can adapt to the current intensity, indicating that the current correction can be further strengthened. If the correlation is less than or equal to the correlation threshold, the current remaining deviation is relatively small, or the user has difficulty adapting to the current intensity, and the correction intensity should be reduced.

[0087] Furthermore, if after two consecutive rounds of adjustments, the absolute value of the rate of change of the residual deviation... If the preset rate is used, the user adaptation coefficient K will be compensated and corrected. If the rate of change of residual deviation If it is negative, then ; If the rate of change of residual deviation If it is positive, then ;in, This is the remaining deviation; The intensity adjustment step size was recalculated based on the corrected user adaptation coefficient. .

[0088] Specifically, the absolute value of the rate of change of the residual deviation > If the preset rate is set, it means that the remaining deviation is changing too quickly.

[0089] If the rate of change of residual deviation If the value is negative and its absolute value is too large, the deviation between the user's action and the standard action is decreasing. However, if the deviation is decreasing too rapidly, K is reduced to decrease the intensity adjustment step size. Corrected intensity adjustment step size Reduce the force to prevent the correction force from decreasing too quickly, which could lead to unstable movements and to avoid overcorrection.

[0090] If the rate of change of residual deviation A positive value with an excessively large absolute value indicates that the deviation between the user's action and the standard action is increasing, and the deviation is increasing too rapidly, requiring an increase in the intensity adjustment step size. Accelerate the correction and push the actions back to the standard.

[0091] To better implement the above method, embodiments of this application also provide a posture sensing device based on a flexible wearable robot, referring to... Figure 2 The attitude sensing device 200 based on a flexible wearable robot includes: The data acquisition module 201 is used to simultaneously collect user standard motion data and real-time motion data. The standard motion data includes the overall motion acceleration data of the wearable part, angular velocity data, deformation resistance change data of the flexible substrate, and real-time ambient temperature data. The initial six-degree-of-freedom attitude data calculation module 202 is used to perform coarse attitude calculation based on the overall motion acceleration data and angular velocity data, and obtain the initial six-degree-of-freedom attitude data through a complementary filtering algorithm. The attitude correction and motion artifact removal module 203 is used to calibrate the initial six-degree-of-freedom attitude data based on the deformation resistance change data, real-time ambient temperature data and the pre-built attitude deformation calibration model, and to complete the attitude drift correction and motion artifact removal to obtain the calibrated six-degree-of-freedom attitude data. The model calibration module 204 is used to extract user-specific motion features and temperature-related features based on the calibrated six-degree-of-freedom attitude data, and dynamically correct the core parameters of the attitude deformation calibration model according to the user-specific motion features and temperature-related features. The attitude data output module 205 is used to input real-time motion data into the corrected attitude deformation calibration model and output real-time attitude data.

[0092] The various variations and specific examples of the methods in the foregoing embodiments are also applicable to the attitude sensing device based on the flexible wearable robot in this embodiment. Through the foregoing detailed description of the attitude sensing method based on the flexible wearable robot, those skilled in the art can clearly understand the implementation method of the attitude sensing device based on the flexible wearable robot in this embodiment. Therefore, for the sake of brevity, it will not be described in detail here.

[0093] This application also provides a computer-readable storage medium storing a computer program. When executed by a processor, the program implements the attitude perception method based on a flexible wearable robot provided in the above embodiments. Through a full-process design of multi-dimensional data synchronous acquisition, complementary filtering coarse calculation, deformation temperature joint calibration, and dynamic model correction, the deformation of the flexible substrate is transformed from an interference source into a calibration basis. Combined with temperature compensation, the attitude drift and motion artifact errors are significantly reduced, the attitude angle perception accuracy is improved, and the model parameters are dynamically corrected based on user-specific features to achieve personalized adaptation.

[0094] In this embodiment, the computer-readable storage medium can be a tangible device that holds and stores instructions used by an instruction execution device. The computer-readable storage medium can be, but is not limited to, an electrical storage device, a magnetic storage device, an optical storage device, an electromagnetic storage device, a semiconductor storage device, or any combination thereof. Specifically, the computer-readable storage medium can be a portable computer disk, a hard disk, a USB flash drive, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), static random access memory (SRAM), portable compact disc read-only memory (CD-ROM), digital multifunction disc (DVD), memory stick, floppy disk, optical disk, magnetic disk, mechanical encoding device, or any combination thereof.

[0095] The computer program in this embodiment includes program code for performing all the aforementioned methods. The program code may include instructions corresponding to the method steps provided in the above embodiments. The computer program can be downloaded from a computer-readable storage medium to various computing / processing devices, or downloaded to an external computer or external storage device via a network (e.g., the Internet, a local area network, a wide area network, and / or a wireless network). The computer program can be executed entirely on the user's computer as a standalone software package.

[0096] The above are all preferred embodiments of this application, and are not intended to limit the scope of protection of this application. Therefore, all equivalent changes made in accordance with the structure, shape and principle of this application should be covered within the scope of protection of this application.

[0097] Additionally, it should be understood that relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. The terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

Claims

1. A pose perception method based on a flexible wearable robot, characterized by, The method comprises the following steps: Synchronously collecting standard motion data and real-time motion data of a user, wherein the standard motion data comprises whole motion acceleration data of a wearing part, angular velocity data, deformation resistance change data of a flexible substrate, and real-time environment temperature data; Based on the whole motion acceleration data and the angular velocity data, coarse attitude calculation is performed through a complementary filtering algorithm to obtain initial six-degree-of-freedom attitude data; Based on the deformation resistance change data, the real-time environment temperature data, and a pre-constructed attitude deformation calibration model, the initial six-degree-of-freedom attitude data is calibrated to complete attitude drift correction and motion artifact elimination, and calibrated six-degree-of-freedom attitude data is obtained; Based on the calibrated six-degree-of-freedom attitude data, user-specific motion features and temperature correlation features are extracted, and core parameters of the attitude deformation calibration model are dynamically corrected according to the user-specific motion features and the temperature correlation features; The real-time motion data is input into the corrected attitude deformation calibration model, and real-time attitude data is output.

2. The method of claim 1, wherein, The calibration of the initial six-degree-of-freedom attitude data based on the deformation resistance change data, the real-time environment temperature data, and the pre-constructed attitude deformation calibration model to complete attitude drift correction and motion artifact elimination comprises: The pre-constructed posture morphing calibration model expression is: ; wherein, is the calibrated attitude angle, is the angle component in the initial six-degree-of-freedom attitude data, and k is a coupling coefficient, is the deformation resistance change rate, i.e., the ratio of the resistance change collected by the strain sensitive grid to the initial resistance, is a temperature compensation term, is the real-time ambient temperature, is the calibration reference temperature; The angle component in the initial six-degree-of-freedom attitude data, the deformation resistance change data, and the real-time environment temperature data are substituted into the model to obtain output calibrated six-degree-of-freedom attitude data; When the attitude angle mutation amplitude in the initial six-degree-of-freedom attitude data or the calibrated six-degree-of-freedom attitude data is greater than an artifact judgment threshold δ, it is determined that there is a motion artifact, and the initial six-degree-of-freedom attitude data or the calibrated six-degree-of-freedom attitude data is eliminated.

3. The method of claim 2, wherein, The user-specific motion features include joint range of motion features, deformation rate features, attitude deformation correlation features, and motion stability features, and the dynamic adjustment of the core parameters of the attitude deformation calibration model according to the user-specific motion features and the temperature correlation features comprises: Based on the joint range of motion features, the maximum bending angle of a user joint is obtained, based on the deformation rate features, the average deformation rate of a user joint is obtained, based on the attitude deformation correlation features, an attitude deformation correlation coefficient is obtained, based on the motion stability features, the attitude angle standard deviation of a repeated motion is obtained, and based on the temperature correlation features, real-time environment temperature and temperature change rate are obtained; Based on the user-specific motion features, the coupling coefficient k, the temperature compensation term ε, and the artifact judgment threshold δ of the attitude deformation calibration model are adjusted through a quantitative correction formula; The correction formula of the coupling coefficient k is: ; wherein, is a general factory coupling coefficient, is a maximum bending angle of a user joint, is a general standard maximum bending angle of a joint, is a user posture deformation correlation coefficient, is a general standard correlation coefficient, is a calibration reference temperature, is a real-time environment temperature, , is a motion feature weight coefficient, is a temperature feature weight coefficient; The modified formula of the temperature compensation term ε is: ; wherein, is a general temperature compensation value, is an average joint deformation rate of the user, is a general standard average deformation rate, is a temperature change rate, is a general standard temperature change rate, is a deformation rate weight coefficient, is a temperature characteristic weight coefficient; The correction formula of the artifact determination threshold δ is: ; wherein, is a general artifact decision threshold, is a standard deviation of the posture angle of the user's repetitive motion, is a general standard motion standard deviation, is a weight coefficient.

4. The method of claim 1, wherein, The whole motion acceleration data and the angular velocity data are collected by an IMU chip, and the deformation resistance change data is collected by a strain-sensitive grid, and the method further comprises: The attitude data is matched with a motion trend mapping library to obtain a current attitude change trend, and a motion complexity level of the user within a preset time in the future is predicted, wherein the motion complexity level is determined by the attitude angle change rate and the number of joints involved; Based on the predicted motion complexity level, a sampling scheme is adjusted in reverse. If the action complexity level is low, the IMU chip sampling precision is reduced, the amplification of the strain sensitive grid signal is lowered, and the IMU chip redundant operation is turned off; If the action complexity level is medium, the IMU chip sampling precision and the amplification of the strain sensitive grid are maintained, and the IMU chip basic operation is activated; If the action complexity level is high, the IMU chip sampling precision is increased, the amplification of the strain sensitive grid signal is increased, and all operation units of the IMU chip are activated and coupled with the preloaded calibration model parameters.

5. The method of claim 1, wherein, The method further comprises: pre-acquiring personalized feature data of the user, and constructing a user adaptation coefficient K based on the personalized feature data; wherein the user adaptation coefficient K is dynamically updated with the user's rehabilitation progress; Obtaining real-time posture data of the target rehabilitation part performing the rehabilitation action and comparing the real-time posture data with a standard rehabilitation action to obtain an initial action deviation value , and combining the user adaptation coefficient K to calculate an initial correction strength , and taking the initial correction strength as a current correction strength; wherein, is a basic strength; driving the flexible execution module to output a correction force according to the current correction intensity; Collect posture feedback data, calculate the remaining deviation of the posture feedback data from a standard rehabilitation action, determine a strength adjustment step based on the association between the remaining deviation and a user adaptation coefficient K Add the current deviation correction strength and the strength adjustment step to obtain an updated deviation correction strength, and use the updated deviation correction strength as a new current deviation correction strength The flexible execution module is repeatedly driven to output a correction force according to the current correction intensity; posture feedback data is collected, a remaining deviation of the posture feedback data from a standard rehabilitation action is calculated, and a strength adjustment step is determined based on an association between the remaining deviation and a user adaptation coefficient K , wherein an updated correction intensity is obtained according to the current correction intensity and the strength adjustment step , and the updated correction intensity is taken as a new current correction intensity, until the remaining deviation is less than a preset threshold, and a final correction intensity is determined.

6. The method of claim 5, wherein, determining the intensity adjustment step length based on the association between the residual deviation and the user adaptation coefficient K obtaining an updated deviation correction intensity according to the current deviation correction intensity and the intensity adjustment step length obtaining an updated deviation correction intensity according to the current deviation correction intensity and the intensity adjustment step length calculating the product of the residual deviation and the user adaptation coefficient K to obtain a correlation; If the correlation relationship > correlation threshold, then the current deviation correction intensity is added to the deviation correction intensity of the previous frame to obtain an updated deviation correction intensity. The step size is adjusted, and the current deviation correction intensity is added to the deviation correction intensity of the previous frame to obtain an updated deviation correction intensity. If the correlation is less than or equal to a correlation threshold, then the current correction strength is adjusted by a step size, and the current correction strength is subtracted from the adjusted correction strength to obtain an updated correction strength. a step size, and the current correction strength is subtracted from the adjusted correction strength to obtain an updated correction strength. wherein i is the number of adjustment steps, i∈[1, n], when i = 1, is the base step.

7. The method of claim 5, wherein, The method further comprises: the personalized characteristic data comprises maximum voluntary contraction force of a target part of the user , maximum fluctuation joint range of motion , and the constructing the user adaptation coefficient K based on the personalized characteristic data comprises: , wherein, is a mean value of maximum voluntary contraction force of a healthy population corresponding to an age stage of the user, is a mean value of maximum fluctuation joint range of motion of a healthy population corresponding to an age stage of the user; real-time compensation of the user adaptation coefficient K, including: If after two consecutive rounds of adjustment, the absolute value of the rate of change of the residual deviation If the preset rate is reached, the user adaptation coefficient K will be compensated and corrected. if the remaining deviation change rate is negative, then ; if the rate of change of the residual deviation is positive, then ; wherein, is the residual deviation; recomputing the strength adjustment step based on the revised user adaptation coefficient .

8. A posture sensing device based on a flexible wearable robot, characterized by, including: a data acquisition module, configured to synchronously acquire user standard action data and real-time action data, the standard action data including overall motion acceleration data of a wearing part, angular velocity data, deformation resistance change data of a flexible substrate, and real-time environmental temperature data; an initial six-degree-of-freedom attitude data solving module, configured to perform coarse attitude solving on the initial six-degree-of-freedom attitude data based on the overall motion acceleration data and the angular velocity data through a complementary filtering algorithm; an attitude correction and motion artifact removal module, configured to calibrate the initial six-degree-of-freedom attitude data based on the deformation resistance change data, the real-time environmental temperature data, and a pre-constructed attitude deformation calibration model, to complete attitude drift correction and motion artifact removal, and obtain calibrated six-degree-of-freedom attitude data; a model correction module, configured to extract user-specific action features and temperature correlation features based on the calibrated six-degree-of-freedom attitude data, and dynamically correct core parameters of the attitude deformation calibration model according to the user-specific action features and the temperature correlation features; an attitude data output module, configured to input the real-time action data into the corrected attitude deformation calibration model, and output real-time attitude data.

9. A flexible wearable robot, characterized by, including: a flexible composite sensing layer, which is sequentially composed of a polyimide substrate, a silver nanowire conductive network, and a PDMS packaging layer from top to bottom; the flexible composite sensing layer is integrated with a three-axis MEMS IMU chip, a strain sensitive grid, and a temperature compensation unit; a data processing module, including an SoC chip integrated with an ARM Cortex-M7 core and an FPGA coprocessor; the flexible composite sensing layer is electrically connected to the data processing module; a flexible execution module, which communicates with the data processing module through a wireless communication module; a memory; At least one computer program, wherein the at least one computer program is stored in the memory and configured to be executed by the data processing module, the at least one computer program is configured to perform the posture perception method based on the flexible wearable robot as claimed in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, A computer program stored in the memory and capable of being loaded and executed by the processor to perform the posture perception method based on the flexible wearable robot as claimed in any one of claims 1 to 7.

Images