Real-time detection method and system for lower limb binding state of exoskeleton robot

Abstract

The invention discloses a real-time detection method and system for the lower limb binding state of an exoskeleton robot. The method comprises the following steps: acquiring a human body gait in a reversing motion state at present, performing feature extraction on the human body gait, and judging a current gait phase through the feature extraction; and then the lower limb binding state of the corresponding gait phase is detected in real time. The system comprises a state observer module, a gait phase recognition module, a binding detection module, an active disturbance rejection controller module and a CMSIS-DSP library module. Through a split-phase dynamic detection strategy, the real-time performance and reliability of binding and falling identification are remarkably improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of robot control, and in particular relates to a real-time detection method and system for the binding status of the lower limbs of an exoskeleton robot. Background Art

[0002] Exoskeletons are wearable assistive devices that enhance or assist the movement of human limbs. They are widely used in rehabilitation medicine, industrial handling, military load-bearing, and other fields. To effectively collaborate with the human body, exoskeletons must be tightly bound to the user's body. Straps, brackets, or custom support structures mechanically connect the robot's joints to the user's joints, enabling movement synchronization and force transmission.

[0003] In practical applications, the quality of the binding directly impacts the exoskeleton's motion control performance and user comfort. If the binding is too loose, the robot's movements may become out of sync with the human's, leading to problems such as force-position decoupling, control instability, and friction slippage. If the binding is too tight, it can easily cause localized compression, obstructed blood circulation, or user discomfort. Furthermore, during dynamic movements such as walking, running, or weight-bearing, the binding status may change over time, potentially becoming loose, misaligned, or shifted, further compromising safety and control accuracy.

[0004] Therefore, real-time monitoring of the binding status has become a key technology for improving the performance of exoskeleton human-machine collaboration. Current research has explored the use of pressure sensors, strain gauges, inertial measurement units, or vision systems to detect binding status, but these still face challenges such as high cost, complex deployment, and susceptibility to interference. Summary of the Invention

[0005] The present invention aims to provide a real-time detection method and system for the lower limb binding status of an exoskeleton robot, thereby improving the reliability and safety of exoskeleton control.

[0006] To achieve the purpose of the present invention, on the one hand, the present invention provides a real-time detection method for the lower limb binding state of an exoskeleton robot, comprising the following steps:

[0007] Step 1: collecting the gait of a human body currently in a commutation motion state and extracting features thereof, and determining the current gait phase by the feature extraction;

[0008] Step 2: When the current gait phase is judged to be in the stance phase, whether the binding has fallen off is determined by the mean and variance strategy of the speed error;

[0009] Step 3: When it is determined that the current gait phase is in the swing phase, whether the binding has fallen off is determined according to the kinematic consistency strategy;

[0010] Step 4: After the above judgment, the real-time detection of the lower limb binding status is completed.

[0011] On the other hand, the present invention also provides a real-time detection system for the lower limb binding status of an exoskeleton robot, comprising the following modules:

[0012] The state observer module is used to low-pass filter the collected raw sensor data, improve the data quality, and generate the standardized input required for gait phase recognition;

[0013] The gait phase recognition module is used to extract gait motion characteristics based on joint angle and angular velocity data, and determine whether the gait is in the stance phase or swing phase based on the joint angle difference and angular velocity difference, providing prior information for subsequent binding state judgment;

[0014] The binding detection module is used to execute corresponding fall-off detection strategies in different gait phases;

[0015] The active disturbance rejection controller module is used to adjust the motor joint speed for anti-disturbance stability during the stance phase binding detection, ensuring that gravity compensation can be provided when the binding falls off and the leg bar rises;

[0016] This CMSIS-DSP library module is used to calculate the mean and variance of joint angular velocity errors during stance phase binding detection. It provides algorithm acceleration and functions for calculating the mean and variance, ensuring high-frequency response requirements for lower limb movements.

[0017] Compared with the existing technology, the significant progress of the present invention lies in: (1) the present invention adopts different strategies in the stance phase and the swing phase respectively, so as to realize the effective identification of the lower limb binding detachment or loose state, thereby improving the reliability and safety of exoskeleton control; (2) the present invention can adapt to the real-time monitoring of the binding status in a dynamic gait environment, and has the advantages of high accuracy, good real-time performance, and no reliance on additional hardware sensors.

[0018] In order to more clearly illustrate the functional characteristics and structural parameters of the present invention, further description is given below with reference to the accompanying drawings and specific embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] The drawings described herein are used to provide a further understanding of the present invention and constitute a part of this application. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings:

[0020] Figure 1 It is a flow chart of the steps of the present invention;

[0021] Figure 2 It is a schematic diagram of the standing phase binding detection of the present invention. DETAILED DESCRIPTION

[0022] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments; based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0023] The present invention provides a real-time detection method for the lower limb binding state of an exoskeleton robot, combined with Figure 1 , including the following steps:

[0024] Step 1: collecting the gait of a human body currently in a commutation motion state and extracting features thereof, and determining the current gait phase by the feature extraction;

[0025] Step 2: When the current gait phase is judged to be in the stance phase, whether the binding has fallen off is determined by the mean and variance strategy of the speed error;

[0026] Step 3: When it is determined that the current gait phase is in the swing phase, whether the binding has fallen off is determined according to the kinematic consistency strategy;

[0027] Step 4: After the above judgment, the real-time detection of the lower limb binding status is completed.

[0028] The commutation motion state in step 1 includes the angle of the leg joint, the angular velocity curve, and the linear acceleration change of the inertial measurement unit (IMU).

[0029] The feature extraction in step 1 is performed by comparing the collected joint angle difference and joint angular velocity difference with their preset thresholds; when the joint angle difference and the joint angular velocity difference are less than or equal to the preset threshold, it is judged that the current gait phase is in the stance phase; when the joint angle difference is greater than the preset threshold and the joint angular velocity difference is positive (i.e., the angle difference shows an expanding trend but has not reached the swing peak), it is judged that the current gait phase is in the swing phase.

[0030] The preset threshold value of the joint angle difference is 0.4 rad, and the preset threshold value of the joint angular velocity difference is 0.5 rad / s.

[0031] Combine Figure 2 , the step 2 specifically includes the following steps:

[0032] Step 2-1, setting the expected angular velocity value, and detecting the actual value of the angular velocity of each joint of the exoskeleton robot once every control cycle (e.g., 10ms);

[0033] Step 2-2: Set a sliding time window T (e.g., 300ms) to observe the joint angular velocity error. The collected real-time data is first low-pass filtered by the state observer, and then the sample value of the velocity error is determined by the expected angular velocity value and the actual value of the angular velocity of each joint:

[0034] ;

[0035] in, is the expected value of angular velocity, is the actual value of the angular velocity of each joint, is the sample value of the velocity error;

[0036] The sliding time window T defines the time range for observing the speed error. The real-time data collected within the window is used to generate sample values ​​of the speed error through calculation. The sliding of the window dynamically updates the sample set.

[0037] Step 2-3: When the sample value reaches the upper limit of the sliding time window, the mean value of the speed error is determined by calculating the mean function arm_mean_f32(e_buf, N, &mean) using the CMSIS-DSP library (a digital signal processing function library based on the Cortex microcontroller software interface standard). , determine the variance of the speed error by calculating the variance function arm_variance_f32(e_buf, N, &var) , where e_buf represents the sample array, N represents the number of sample data, mean represents the calculated mean, and var represents the calculated variance;

[0038] Step 2-4: Obtain the mean value of the speed error and variance After that, it is determined whether the current mean and variance meet certain conditions, and whether the binding has fallen off.

[0039] The conditions satisfied in steps 2-4 are specifically shown in the following formula:

[0040] ;

[0041] in, is the lower bound of the mean velocity error, is the upper limit of the mean velocity error, is the threshold of velocity error variance, is the mean velocity error of the data in the sliding time window, is the velocity error variance of the data within the sliding time window;

[0042] When the speed error mean is within the range of ±0.05 rad / s and the speed error variance is less than 0.03 rad2 / s2, the current binding is determined to be at risk of falling off or slipping, and an alarm is issued and the exoskeleton robot is disabled (the device is forced to power off);

[0043] When the speed error mean exceeds a preset interval or the speed error variance exceeds a preset threshold, it is determined that the current binding is in a normal state, and the binding state is continuously detected.

[0044] The step 3 specifically includes the following steps:

[0045] Step 3-1: Collect motion features, including the angle and angular velocity of the stance leg and swing leg when switching, and record the gait reversal angle of N consecutive gait cycles (e.g., 5) and commutation angular velocity ;

[0046] Step 3-2: Based on the gait commutation angle, a theoretical commutation angle of the current gait cycle is estimated by linear fitting;

[0047] Step 3-3: Based on the commutation angular velocity, estimate the range of the joint angular velocity when the theoretical commutation angle is reached by using the deviation. ,in is the minimum value of the commutation angular velocity over multiple consecutive cycles, is the maximum value of the commutation angular velocity over multiple consecutive cycles;

[0048] Step 3-4: When the current joint angular velocity is measured in real time After reaching the theoretical commutation angle, it is determined whether the joint angular velocity meets certain conditions, and whether the binding has fallen off:

[0049] The theoretical tangential angle of step 3-2 , as shown in the following formula:

[0050] ;

[0051] in, represents the parameters of the linear fit, , represents the gait commutation angle of the latest N cycles, represents the initial gait turning angle (usually 0.2rad).

[0052] The conditions satisfied in steps 3-4 are specifically shown in the following formula:

[0053] ;

[0054] in, This is the speed redundancy tolerance. In actual applications, due to factors such as different response times of different motors, different control filter parameter settings, or signal processing delays, the actual joint angular velocity may temporarily exceed the commutation speed range obtained by fitting the previous N cycles even when the binding is normal. This parameter is mainly used to avoid misjudgment. The value of the tolerance δ is set based on the speed fluctuation data, filter bandwidth settings, and historical maximum deviation during the system debugging process, and is usually set between 0.1-0.3 rad / s. is the actual value of the angular velocity of each joint;

[0055] If the above formula is satisfied, it is determined that the binding has fallen off or slipped, and an alarm is issued and the exoskeleton robot is disabled for protection; otherwise, the process returns to step 3-1 to continue binding detection.

[0056] The present invention provides a real-time detection system for the lower limb binding status of an exoskeleton robot, comprising the following modules:

[0057] The state observer module is used to low-pass filter the collected raw sensor data (such as joint angles and angular velocities), improve the data quality, and generate the standardized input required for gait phase recognition;

[0058] The gait phase recognition module is used to extract gait motion characteristics based on joint angle and angular velocity data, and determine whether the gait is in the stance phase or swing phase based on the joint angle difference and angular velocity difference, providing prior information for subsequent binding state judgment;

[0059] The binding detection module is used to implement corresponding detachment detection strategies in different gait phases. For example, in the stance phase, it determines whether there is loose binding and detachment based on the mean and variance of the joint angular velocity error. In the swing phase, it determines whether there is loose binding and detachment based on whether the current commutation angular velocity deviates from the angular velocity maximum range of the previous cycles.

[0060] An active disturbance rejection controller (ADRC) module is used to perform anti-disturbance stability adjustment on the motor joint velocity during stance phase binding detection, ensuring gravity compensation when the leg bar rises due to a binding detachment, reducing velocity error and thus enhancing the robustness of the binding detection function. When the exoskeleton robot is in the stance phase of its gait, the active disturbance rejection controller (ADRC) module is activated to perform feedforward compensation and disturbance rejection adjustment on the target joint angular velocity. The ADRC module estimates equivalent disturbances (such as gravity and friction) based on the real-time system state and performs online compensation on the control input to improve the accuracy and responsiveness of joint velocity control.

[0061] This CMSIS-DSP library module is used to calculate the mean and variance of joint angular velocity errors during stance phase binding detection. It provides algorithm acceleration and functions for calculating the mean and variance, ensuring high-frequency response requirements for lower limb movements.

[0062] It should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that includes a list of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, article, or apparatus.

[0063] 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 real-time detection method for the lower limb binding status of an exoskeleton robot, characterized in that: The following steps are involved: Step 1: collecting the gait of a human body currently in a commutation motion state and extracting features thereof, and determining the current gait phase by the feature extraction; Step 2: When the current gait phase is judged to be in the stance phase, whether the binding has fallen off is determined by the mean and variance strategy of the speed error; Step 3: When it is determined that the current gait phase is in the swing phase, whether the binding has fallen off is determined according to the kinematic consistency strategy; Step 4: After the above judgment, the real-time detection of the lower limb binding status is completed.

2. The method for real-time detection of the lower limb binding status of an exoskeleton robot according to claim 1, characterized in that: The commutation motion state in step 1 includes the angle of the leg joint, the angular velocity curve, and the linear acceleration change of the inertial measurement unit.

3. The real-time detection method for the lower limb binding status of an exoskeleton robot according to claim 2, characterized in that: The feature extraction in step 1 is achieved by comparing the collected joint angle difference and joint angular velocity difference with their preset thresholds; when the joint angle difference and the joint angular velocity difference are less than or equal to the preset threshold, it is judged that the current gait phase is in the stance phase; when the joint angle difference is greater than the preset threshold and the joint angular velocity difference is a positive value, it is judged that the current gait phase is in the swing phase.

4. The method for real-time detection of the lower limb binding status of an exoskeleton robot according to claim 3, characterized in that: The step 2 specifically includes the following steps: Step 2-1, setting the expected angular velocity value, and detecting the actual value of the angular velocity of each joint of the exoskeleton robot once every control cycle; Step 2-2, setting a sliding time window T to observe the error of the joint angular velocity, first low-pass filtering the collected real-time data through a state observer, and then determining the sample value of the velocity error by using the expected angular velocity value and the actual value of the angular velocity of each joint; Step 2-3: When the sample value reaches the time upper limit set by the sliding time window, the mean of the speed error is determined by calculating the mean function using the CMSIS-DSP library, and the variance of the speed error is determined by calculating the variance function; Step 2-4: After obtaining the mean and variance of the speed error, determine whether the current mean and variance meet certain conditions, and use this to determine whether the binding has fallen off.

5. The method for real-time detection of the lower limb binding status of an exoskeleton robot according to claim 4, characterized in that: The conditions satisfied in steps 2-4 are specifically shown in the following formula: ； in, is the lower bound of the mean velocity error, is the upper limit of the mean velocity error, is the threshold of velocity error variance, is the mean velocity error of the data in the sliding time window, is the velocity error variance of the data within the sliding time window; When the speed error mean is within the range of ±0.05 rad / s and the speed error variance is lower than 0.03 rad2 / s2, the current binding is determined to be in a risk state of falling off or slipping, and an alarm is issued and the exoskeleton robot is disabled for protection; When the speed error mean exceeds a preset interval or the speed error variance exceeds a preset threshold, it is determined that the current binding is in a normal state, and the binding state is continuously detected.

6. The method for real-time detection of the lower limb binding status of an exoskeleton robot according to claim 3, characterized in that: The step 3 specifically includes the following steps: Step 3-1: Collect motion characteristics, including the angle and angular velocity when switching between the stance leg and the swing leg, and record the gait reversal angle and reversal angular velocity for N consecutive gait cycles; Step 3-2: Based on the gait commutation angle, a theoretical commutation angle of the current gait cycle is estimated by linear fitting; Step 3-3: Based on the commutation angular velocity, estimate the range of the joint angular velocity when the theoretical commutation angle is reached by using the deviation. ,in is the minimum value of the commutation angular velocity over multiple consecutive cycles, is the maximum value of the commutation angular velocity over multiple consecutive cycles; Step 3-4: When the current joint angular velocity measured in real time reaches the theoretical commutation angle, it is determined whether the joint angular velocity meets certain conditions, and whether the binding has fallen off.

7. The method for real-time detection of the lower limb binding status of an exoskeleton robot according to claim 6, characterized in that: The theoretical tangential angle of step 3-2 , as shown in the following formula: ； in, represents the parameters of the linear fit, , represents the gait commutation angle of the latest N cycles, represents the initial gait turning angle.

8. The method for real-time detection of the lower limb binding status of an exoskeleton robot according to claim 6, characterized in that: The conditions satisfied in steps 3-4 are specifically shown in the following formula: ； in, is the speed redundancy tolerance; is the actual value of the angular velocity of each joint; If the above formula is satisfied, it is determined that the binding has fallen off or slipped, and an alarm is issued and the exoskeleton robot is disabled for protection; otherwise, the process returns to step 3-1 to continue binding detection.

9. The method for real-time detection of the lower limb binding status of an exoskeleton robot according to claim 3, characterized in that: The preset threshold value of the joint angle difference is 0.4 rad, and the preset threshold value of the joint angular velocity difference is 0.5 rad / s.

10. A real-time detection system for the lower limb binding status of an exoskeleton robot, used to implement the method described in any one of claims 1 to 9, characterized in that: Includes the following modules: The state observer module is used to low-pass filter the collected raw sensor data, improve the data quality, and generate the standardized input required for gait phase recognition; The gait phase recognition module is used to extract gait motion characteristics based on joint angle and angular velocity data, and determine whether the gait is in the stance phase or swing phase based on the joint angle difference and angular velocity difference, providing prior information for subsequent binding state judgment; The binding detection module is used to execute corresponding fall-off detection strategies in different gait phases; The active disturbance rejection controller module is used to adjust the motor joint speed for anti-disturbance stability during the stance phase binding detection, ensuring that gravity compensation can be provided when the binding falls off and the leg bar rises; This CMSIS-DSP library module is used to calculate the mean and variance of joint angular velocity errors during stance phase binding detection. It provides algorithm acceleration and functions for calculating the mean and variance, ensuring that the high-frequency response requirements of lower limb movements are met.

Images