Exoskeleton power-assisted control method based on force feedback
By using motor encoder position signals and PID control strategies, combined with angular velocity-angle composite criteria and state machine FSM, the human-machine interaction of the exoskeleton robot was improved, the problem of insufficient system flexibility and adaptability was solved, and safe and efficient exoskeleton control was achieved.
Patent Information
- Application Number
- CN202511615121.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-11-06
- Publication Date
- 2026-02-24
AI Technical Summary
Existing exoskeleton robots have poor human-computer interaction, insufficient system flexibility and adaptability, and are difficult to adapt to different movement patterns and user needs.
Using the position signal of the motor encoder as the core reference signal for force feedback, and combining PID control strategy and dual guarantee mechanism, a motion intention prediction algorithm based on angular velocity-angle composite criterion is designed. The PID algorithm calculates the precise control of joint angle and torque, and the state machine FSM is used to control the movement of the exoskeleton knee joint.
It achieves a safe, efficient, and comfortable human-machine integration experience, reduces system costs, avoids the risk of sports injuries, and improves the system's flexibility and adaptability.
Smart Images
Figure CN121552337A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to force feedback systems, and more specifically to a force feedback-based exoskeleton assistive control method. Background Technology
[0002] The rapid advancement of science and technology, particularly the rise of exoskeleton robot technology, marks a significant leap forward in the field of human-computer interaction. However, since exoskeleton robots and their wearers interact physically through force at each point of contact, the effectiveness of this interaction largely depends on the controller design. To improve the human-computer interaction effect, a general and modular exoskeleton control framework is proposed, aiming to enhance the system's flexibility and applicability. This framework inherits the core structure of force feedback control and optimizes the reference generation mechanism and force control strategy based on practical application scenarios. By introducing a dynamic adjustment mechanism, the system can better adapt to different motion modes and user needs, thereby achieving more natural and efficient human-computer collaboration. Summary of the Invention
[0003] The technical problem to be solved by the present invention is to provide an exoskeleton-assisted control method based on force feedback, which can effectively improve the human-computer interaction effect, enhance the flexibility and adaptability of the system, and better adapt to different movement modes.
[0004] To address the aforementioned problems, this invention provides an exoskeleton-assisted control method based on force feedback, characterized in that the control method includes: The position signal of the motor encoder is used as the core reference signal for force feedback. Combined with PID control strategy and dual guarantee mechanism, a force feedback system is constructed and a motion intention prediction algorithm based on angular velocity-angle composite criterion is designed. The position signal generated in real time by the motor encoder is obtained. Based on the position signal, the signal is processed to eliminate the boundary error of the angle, and the real-time angle and real-time angular velocity of the exoskeleton joint are calculated. By combining the PID algorithm as a closed-loop control method based on the error feedback principle, the torque control signal for driving the motor is calculated by the PID controller, thus achieving precise control of joint angle and torque. The PID parameter tuning method employed involves manually adjusting the proportional, integral, and derivative parameters of the PID controller sequentially through trial and error, while closely observing the system response to optimize control effectiveness and determine the appropriate parameters. This is to achieve optimal control of the exoskeleton; among which, motor current detection indirectly calculates the interaction force between the exoskeleton and the human body by monitoring the current change of the drive motor.
[0005] The angular velocity-angle composite criterion requires both the angular velocity trigger condition and the angle auxiliary condition to be met in order to predict motion intent.
[0006] The angular velocity triggering condition requires a real-time comparison of the angular velocity with a preset angular velocity threshold, and the angle must be within the safe range set for joint movement. Based on the comparison result, the user's movement intention is determined, and a corresponding movement mode control command is generated. If the angle exceeds the safe range, forced braking will be triggered.
[0007] Determining the motion intent includes triggering a nonlinear torque decay control command if the absolute value of the real-time angular velocity is greater than a first high-speed threshold. If the absolute value of the real-time angular velocity is greater than the first low-speed threshold and less than or equal to the high-speed threshold, then a linear PID assist control command is triggered. If the absolute value of the real-time angular velocity is less than or equal to the first low-speed threshold, the standby mode control command is triggered.
[0008] The system enters standby mode, the motor output is turned off, the integral term in the PID is reset to zero to prevent the integral from accumulating, and it waits for a valid signal that can trigger movement. If the user starts to move, and the angular velocity is detected to exceed the predetermined threshold, the system enters either the knee extension or knee flexion state based on the direction, and switches to the corresponding state to implement motor control accordingly. At the same time, it completes the shortest path planning with the help of modular arithmetic and conditional judgment to conform to the human body's movement.
[0009] The knee joint extension control system continuously calculates the error between the current angle and the target angle, introducing an inverse error compensation strategy to suppress potential overshoot. It updates the target angle to the current position to prevent system oscillations and overshoot. If the angle error is less than the preset tolerance, the system automatically returns to an idle state.
[0010] The control process for knee flexion is basically the same as that for knee extension. However, when setting the target angle, negative offset needs to be considered. If the calculated result is less than 0°, it is mapped to a 360° range using modular calculation to ensure the continuity and appropriateness of the angle representation and form a complete action loop.
[0011] This method uses a state machine (FSM) in three states to control the movement of the exoskeleton knee joint, monitors the joint's angular velocity in real time, and completes the switching between different movement modes. The state machine includes three core states: idle, knee extension, and knee flexion, which correspond to different movement intentions and control strategies.
[0012] In the first dynamic response layer, angular velocity serves as the core parameter for determining the user's motion intention. By setting different thresholds, linear assist and nonlinear torque decay control are triggered respectively. In the second static protection layer, the main function of the angle parameter is to limit the range of motion and calibrate the steady state. A safe range of motion is pre-set for the designed knee joint. The degree of completion of the movement is judged by the deviation between the target angle and the actual angle. If the deviation is less than the set threshold, torque decay control is implemented to enter standby mode.
[0013] Signal processing includes pulse acquisition, angle calculation, and angular velocity calculation. The angular velocity is calculated in real time using the differential method based on the interrupt function, and jump correction is performed to eliminate the boundary error of the angle. If the original angular velocity value calculated from the continuous angles exceeds ±180° / ∆t, it is determined that a circular jump has occurred. This jump is compensated by adding or subtracting 360° / ∆t to obtain an accurate angular velocity value that reflects the real physical motion.
[0014] The beneficial effects of this invention are: This invention achieves a safe, efficient, and comfortable human-computer interaction experience by designing a motion intention prediction algorithm based on a composite angular velocity-angle criterion and combining it with a PID algorithm. Compared to traditional solutions using a single angle criterion, which suffer from latency issues, this invention provides redundant protection for the wearer from both dynamic and static dimensions, fundamentally eliminating the risk of sports injuries.
[0015] By using the motor encoder position signal as the core reference signal for force feedback, this technology successfully replaces high-cost force sensors, torque sensors, or electromyography (EMG) sensors, significantly reducing system cost and complexity. Employing modular arithmetic and jump correction algorithms, the jump problem from 359° to 0° is resolved, ensuring the continuity of angle and angular velocity calculations and thus guaranteeing smooth control. Its significant advantages of low cost, high safety, fast response, and high precision have greatly promoted the practical application and industrialization of lower limb exoskeleton technology in rehabilitation medicine and industrial assistive technology. Attached Figure Description
[0016] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings: Figure 1 This is a schematic diagram illustrating a specific implementation of the exoskeleton-assisted control method based on force feedback proposed in this invention.
[0017] Figure 2 This invention presents a schematic diagram of a dual-protection mechanism for an exoskeleton-based force feedback-assisted control method.
[0018] Figure 3 This invention presents a schematic diagram of a motion intent determination structure for an exoskeleton-assisted control method based on force feedback. Detailed Implementation
[0019] The present invention will now be described in further detail with reference to the accompanying drawings. These drawings are simplified schematic diagrams, illustrating only the basic structure of the invention, and therefore only show the components relevant to the invention.
[0020] refer to Figure 1 A force feedback-based exoskeleton assistive control method is described. The control method includes: using the position signal of the motor encoder as the core reference signal for force feedback, and combining PID control strategy and dual guarantee mechanism to construct a force feedback system, and designing a motion intention prediction algorithm based on angular velocity-angle composite criterion. The position signal generated in real time by the motor encoder is acquired. Based on the position signal, the signal is processed to eliminate the boundary error of the angle, and the real-time angle and real-time angular velocity of the exoskeleton joint are calculated. The signal processing includes pulse acquisition, angle calculation and angular velocity calculation. The angular velocity is calculated in real time using the differential method with the help of the interrupt function and jump correction is performed to eliminate the boundary error of the angle. In this embodiment, the fmod function is used to calculate the remainder after dividing two floating-point numbers; N is the cumulative number of pulses; the encoder resolution is 408. The real-time angle of the joint is obtained and then processed into a ring shape. A difference formula is then used. The real-time angular velocity is obtained, and a jump correction algorithm is applied. If the original angular velocity value calculated from the continuous angles exceeds ±180° / ∆t, it is determined that a circular jump has occurred. The jump of this circle is compensated by adding or subtracting 360° / ∆t to obtain an accurate angular velocity value that reflects the real physical motion, thus ensuring the accuracy of the data.
[0021] Table 1. Correction Processing Table
[0022] As shown in Table 1, when the actual value is detected to be more than 180° away from the target value, it is judged as a circular jump and the direction of the angular velocity is corrected to avoid errors in the calculation process.
[0023] By combining the PID algorithm as a closed-loop control method based on the error feedback principle, the torque control signal for driving the motor is calculated by the PID controller, thus achieving precise control of joint angle and torque. The PID parameter tuning method based on empirical observation involves manually adjusting the proportional, integral, and derivative parameters of the PID controller sequentially through trial and error, while closely observing the system response to optimize control effectiveness. The core of this method lies in using an intuitive "adjust-observe-correct" cycle to ultimately determine the parameters. To achieve optimal control of the exoskeleton.
[0024] Table 2. Experimental Results of PID Parameters
[0025] The final parameter set confirmed by the experiment It performs well in many aspects, as shown in Table 2. Compared with pure proportional control, this combination reduces steady-state error and weakens overshoot. Although increasing the integral term can improve steady-state error to a certain extent, it will weaken system stability and cause motor jitter and oscillation. The introduction of the derivative term significantly reduces the system oscillation index and significantly improves the smoothness of motion.
[0026] Among them, motor current detection indirectly calculates the interaction force between the exoskeleton and the human body by monitoring the current change of the drive motor. This makes it highly real-time, but it is also susceptible to the nonlinear characteristics of the motor, such as friction and temperature drift, which leads to increased noise. Complex filtering algorithms are needed to improve the signal-to-noise ratio. Moreover, in assistive control, motor current detection has difficulty distinguishing between the active force of the human body and environmental resistance. In contrast, encoder signals provide direct position feedback, support closed-loop control of joint angles, have high angle resolution, and are less affected by electromagnetic noise, making them very suitable for high-precision joint control and complex wearable environments.
[0027] The angular velocity-angle composite criterion requires both the angular velocity trigger condition and the angle auxiliary condition to be met in order to predict motion intent.
[0028] This composite criterion gives the angle condition the highest priority. Even if the angular velocity strongly indicates that the user wants to move, such as when the user generates an extremely high angular velocity due to a fall, as long as the angle exceeds the safe range, the system will immediately disable the assist, prioritizing safety and reflecting the design principle of "safety over responsiveness".
[0029] Table 3. Range of motion of joints in the human lower limbs
[0030] In this embodiment, based on the range of motion of the knee joint in Table 3, the angular velocity triggering condition requires a real-time comparison of the angular velocity with a preset angular velocity threshold, and the knee joint angle must be within the safe range of 0° to 130°. The user's movement intention is determined based on the comparison result, and a corresponding movement mode control command is generated. If the angle exceeds the safe range, forced braking will be triggered.
[0031] Determining the motion intent includes triggering a nonlinear torque decay control command if the absolute value of the real-time angular velocity is greater than a first high-speed threshold. If the absolute value of the real-time angular velocity is greater than the first low-speed threshold and less than or equal to the high-speed threshold, then a linear PID assist control command is triggered. If the absolute value of the real-time angular velocity is less than or equal to the first low-speed threshold, the standby mode control command is triggered.
[0032] The system enters standby mode, the motor output is turned off, the integral term in the PID is reset to zero to prevent the integral from accumulating, and it waits for a valid signal that can trigger movement. If the user starts to move, and the angular velocity is detected to exceed the predetermined threshold, the system enters either the knee extension or knee flexion state based on the direction, and switches to the corresponding state to implement motor control accordingly. At the same time, it completes the shortest path planning with the help of modular arithmetic and conditional judgment to conform to the human body's movement.
[0033] The system implements knee joint extension control, continuously calculating the error between the current angle and the target angle. An inverse error compensation strategy is introduced to suppress potential overshoot, updating the target angle to the current position to prevent system oscillations and overshoot. If the angle error is less than the preset tolerance, the system automatically returns to an idle state. The control process for knee flexion is basically the same as that for knee extension. However, when setting the target angle, negative offset needs to be considered. If the calculated result is less than 0°, it is mapped to a 360° range using modular calculation to ensure the continuity and appropriateness of the angle representation and form a complete action loop.
[0034] This method uses a state machine (FSM) in three states to control the movement of the exoskeleton knee joint, monitors the joint's angular velocity in real time, and completes the switching between different movement modes. The state machine includes three core states: idle, knee extension, and knee flexion, which correspond to different movement intentions and control strategies.
[0035] The control process employs a dual-guarantee mechanism. In the first dynamic response layer, angular velocity serves as the core parameter for determining the user's movement intention. Different thresholds are set to trigger linear assist and nonlinear torque decay control, respectively. In the second static protection layer, the angle parameter primarily functions to limit the range of motion and calibrate steady-state. The angular velocity layer ensures response speed, quickly keeping up with the user's movement intention. The angle layer ensures ultimate safety by establishing an insurmountable physical barrier. Even if the dynamic layer misjudges or fails, the static layer still provides protection.
[0036] At the same time, a safe range of motion is set in advance for the designed knee joint. The degree of completion of the movement is judged by the deviation between the target angle and the actual angle. If the deviation is less than the set threshold, torque decay control is performed to enter the standby state. By utilizing a microcontroller to achieve a complete closed loop of angle and angular velocity signal acquisition, state machine decision-making, PID control, and motor drive, it supports both knee extension and knee flexion assist modes. At the same time, it integrates OLED display and serial communication interface to improve debuggability, providing a reliable and flexible technical solution for exoskeleton force feedback control.
[0037] In this embodiment, the encoder position signal is detected in real time, and the instantaneous angular velocity is calculated using the differential method through an interrupt function. Jump correction is then used to eliminate angular boundary errors. The entire process follows a dual protection mechanism prioritizing safety over responsiveness, taking into account both dynamic response and static protection. Subsequently, the system uses a composite angular velocity-angle criterion to predict the user's motion intention, completing the switching between different motion modes. The PID controller dynamically calculates a precise assist torque based on the angular error, and finally, the assist is delivered through the motor output, forming a closed-loop control from signal sensing to power output.
[0038] Using an encoder as the force feedback signal source achieves a balance between high performance and low cost. By utilizing angular velocity as the primary criterion, it can capture intent even when the motion involves minute displacements, fundamentally solving the inherent latency problem of position-based judgment methods. Simultaneously, a dual-insurance mechanism ensures the system's final behavior is safe under all circumstances, greatly enhancing user trust. It exhibits strong anti-interference capabilities and stable performance across different users and scenarios.
[0039] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.
Claims
1. A force feedback-based exoskeleton assistive control method, characterized in that, The control method includes: The position signal of the motor encoder is used as the core reference signal for force feedback. Combined with PID control strategy and dual guarantee mechanism, a force feedback system is constructed and a motion intention prediction algorithm based on angular velocity-angle composite criterion is designed. The position signal generated in real time by the motor encoder is obtained. Based on the position signal, the signal is processed to eliminate the boundary error of the angle, and the real-time angle and real-time angular velocity of the exoskeleton joint are calculated. By combining the PID algorithm as a closed-loop control method based on the error feedback principle, the torque control signal for driving the motor is calculated by the PID controller, thus achieving precise control of joint angle and torque. The PID parameter tuning method employed involves manually adjusting the proportional, integral, and derivative parameters of the PID controller sequentially through trial and error, while closely observing the system response to optimize control effectiveness and determine the appropriate parameters. This is to achieve optimal control of the exoskeleton; among which, motor current detection indirectly calculates the interaction force between the exoskeleton and the human body by monitoring the current change of the drive motor.
2. The exoskeleton assistive control method based on force feedback according to claim 1, characterized in that... The angular velocity-angle composite criterion requires both the angular velocity trigger condition and the angle auxiliary condition to be met in order to predict motion intent.
3. The exoskeleton assistive control method based on force feedback according to claim 2, characterized in that... The angular velocity triggering condition requires a real-time comparison between the angular velocity and a preset angular velocity threshold, and the angle must be within the safe range set for joint movement. Based on the comparison result, the user's movement intention is determined, and corresponding movement mode control commands are generated. If the angle exceeds the safe range, forced braking will be triggered.
4. The exoskeleton assistive control method based on force feedback according to claim 3, characterized in that... Determining the motion intent includes triggering a nonlinear torque decay control command if the absolute value of the real-time angular velocity is greater than a first high-speed threshold. If the absolute value of the real-time angular velocity is greater than the first low-speed threshold and less than or equal to the high-speed threshold, then a linear PID assist control command is triggered. If the absolute value of the real-time angular velocity is less than or equal to the first low-speed threshold, the standby mode control command is triggered.
5. The exoskeleton assistive control method based on force feedback according to claim 4, characterized in that... The system enters standby mode, the motor output is turned off, the integral term in the PID is reset to zero to prevent the integral from accumulating, and it waits for a valid signal that can trigger movement. If the user starts to move and the angular velocity is detected to exceed the predetermined threshold, the system enters one of the two states of knee extension or knee flexion based on the direction, and switches to the corresponding state to implement motor control accordingly. At the same time, it completes the shortest path planning with the help of modular arithmetic and conditional judgment to conform to the human body's movement.
6. The exoskeleton assistive control method based on force feedback according to claim 4, characterized in that... The knee joint extension control system continuously calculates the error between the current angle and the target angle, introducing an inverse error compensation strategy to suppress potential overshoot. It updates the target angle to the current position to prevent system oscillations and overshoot. If the angle error is less than the preset tolerance, the system automatically returns to an idle state.
7. The exoskeleton assistive control method based on force feedback according to claim 4, characterized in that... The control process for knee flexion is basically the same as that for knee extension. However, when setting the target angle, negative offset needs to be considered. If the calculated result is less than 0°, it is mapped to a 360° range using modular calculation to ensure the continuity and appropriateness of the angle representation and form a complete action loop.
8. The exoskeleton assistive control method based on force feedback according to claim 1, characterized in that... This method uses a state machine (FSM) in three states to control the movement of the exoskeleton knee joint, monitors the joint's angular velocity in real time, and completes the switching between different movement modes. The state machine includes three core states: idle, knee extension, and knee flexion, which correspond to different movement intentions and control strategies.
9. The exoskeleton assistive control method based on force feedback according to claim 1, characterized in that... In the first dynamic response layer, angular velocity serves as the core parameter for determining the user's motion intention. By setting different thresholds, linear assist and nonlinear torque decay control are triggered respectively. In the second static protection layer, the main function of the angle parameter is to limit the range of motion and calibrate the steady state. A safe range of motion is pre-set for the designed knee joint. The degree of completion of the movement is judged by the deviation between the target angle and the actual angle. If the deviation is less than the set threshold, torque decay control is implemented to enter standby mode.
10. The exoskeleton assistive control method based on force feedback according to claim 1, characterized in that... Signal processing includes pulse acquisition, angle calculation, and angular velocity calculation. The angular velocity is calculated in real time using the differential method based on the interrupt function, and jump correction is performed to eliminate the boundary error of the angle. If the original angular velocity value calculated from the continuous angles exceeds ±180° / ∆t, it is determined that a circular jump has occurred. This jump is compensated by adding or subtracting 360° / ∆t to obtain an accurate angular velocity value that reflects the real physical motion.