Force Feedback Nonlinear Active Disturbance Rejection Controller and Method Based on Particle Swarm Optimization

Abstract

This invention discloses a force feedback nonlinear active disturbance rejection controller and method based on particle swarm optimization, comprising: a signal input module, a tracking differentiator, a feedback control law, an extended state observer, a force feedback system, and a particle swarm algorithm module; the control method includes: step one, building a model; step two, determining the spatial expression; step three, transient processing of displacement signals; step four, processing feedback control signals; step five, processing extended state observation signals; and step six, compensating for the control quantity feedback system. This invention uses the extended state observer for real-time estimation to compensate for the total disturbance received by the force feedback system during operation. Furthermore, the designed nonlinear active disturbance rejection control technology can perform force feedback control without relying on a specific system model, exhibiting strong adaptability and robustness. The particle swarm optimization algorithm can automatically tune parameters, improving the design efficiency of the nonlinear active disturbance rejection controller and optimizing control performance.

Description

Technical Field

[0001] This invention relates to the field of force feedback manipulator control technology, specifically to a force feedback nonlinear active disturbance rejection controller and method based on particle swarm optimization. Background Technology

[0002] In cases of neurological disorders such as stroke, spinal cord injury, and myasthenia gravis, or arm injuries, many patients require long-term self-rehabilitation. However, traditional rehabilitation methods often have limitations. For example, traditional physical therapy and drug treatment may not be able to achieve precise and personalized rehabilitation. Force feedback robotic arms are rehabilitation devices that use robotics technology to provide patients with assisted training. By accurately sensing, imitating, and controlling hand movements, they provide patients with personalized rehabilitation plans, thereby improving rehabilitation outcomes. By using force feedback robotic arms, patients can rehabilitate more conveniently, enhancing their confidence and self-assurance in rehabilitation. Furthermore, force feedback robotic arms can also promote the development of robotics technology, enabling the creation of intelligent and personalized rehabilitation plans. Therefore, force feedback robotic arms have broad application prospects.

[0003] The existing control technology for force feedback manipulators has the following drawbacks: First, the existing control technology lacks strong adaptability and depends on specific system models. Second, it cannot provide real-time feedback compensation control and is easily affected by interference. Third, the existing particle swarm optimization algorithm cannot automatically tune parameters, and the parameters of nonlinear active disturbance rejection control are too numerous and complex, resulting in low design efficiency. Summary of the Invention

[0004] The purpose of this invention is to provide a force feedback nonlinear active disturbance rejection controller and method based on particle swarm optimization to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a force feedback nonlinear active disturbance rejection controller based on particle swarm optimization, comprising a signal input module, a tracking differentiator, a feedback control law, an extended state observer, a force feedback system, and a particle swarm algorithm module. The output terminals of the signal input module and the particle swarm algorithm module are electrically connected to the tracking differentiator. The output terminals of the tracking differentiator and the extended state observer are electrically connected to the feedback control law. The output terminal of the feedback control law is electrically connected to the force feedback system and the extended state observer. The output terminal of the force feedback system is electrically connected to the particle swarm algorithm module and the extended state observer, respectively.

[0006] The control method based on particle swarm optimization force feedback nonlinear active disturbance rejection controller includes the following steps: Step 1, model building; Step 2, determining the spatial expression; Step 3, transient processing of displacement signals; Step 4, processing of feedback control signals; Step 5, processing of extended state observation signals; and Step 6, compensation of control quantity feedback system.

[0007] In step one above, the position variables of the generated particles in the search space are obtained through particle swarm optimization. and the velocity of particles Build kinematic and dynamic models of force feedback manipulators;

[0008] In step two above, the control output is obtained by optimizing the parameters using the particle swarm optimization method. Determine the state-space expression of the force feedback manipulator;

[0009] In step three above, the signal input module obtains the displacement signal. The input is fed into the tracking differentiator, and the output is... The signal is the input displacement signal. The signal after transition The signal is a displacement signal. The differential signal;

[0010] In step four above, after tracking the differentiator transition signal... Signals and The signals are respectively compared with the output of the extended state observer. and Subtraction yields the input to the feedback control law. and The feedback control law is implemented through a nonlinear combination method, and the calculated output is: ;

[0011] In step five above, the original state variable is expanded in the extended state observer, transforming the total disturbance of the sum of external and internal disturbances into a new state variable. The input signal of the extended state observer is... and The output is , and ;

[0012] In step six above, the compensated control quantity is applied to the force feedback manipulator system to achieve precise control.

[0013] Preferably, in step one, the kinematic model includes a theoretical model of the spring tension and a displacement model of the motor driven by the track, with the following formulas:

[0014]

[0015]

[0016] By combining the two equations and differentiating them separately, we can obtain the speed control quantity for the DC motor:

[0017]

[0018] The dynamic model of the force feedback manipulator is described below.

[0019]

[0020] Preferably, in step two, the state variable is taken. , Therefore, the state equation of the force feedback system is:

[0021]

[0022] The state-space expression of the force feedback system is as follows.

[0023]

[0024]

[0025] Preferably, in step three, the system state of the output port of the differentiator is tracked. and The discrete formula is as follows:

[0026]

[0027] The formula for the tracking differentiator contains a core algorithmic component: the steepest control function. This allows the entire tracking differentiator to quickly and smoothly track changes in the input signal. The tracking formula is as follows:

[0028]

[0029] The tracking differentiator includes the tracking velocity factor as a parameter. and filter factor .

[0030] Preferably, in step four, the expression for the feedback control law is as follows:

[0031]

[0032] in It is a nonlinear function:

[0033]

[0034] The parameters included in a feedback control law are: , , , and .

[0035] Preferably, in step five, a nonlinear extended state observer is established for the force feedback manipulator system as follows:

[0036]

[0037] The discrete form of the extended observer is represented by the following formula:

[0038]

[0039] The parameters included in the extended observer are: , , , , and ,in and You can choose any value; taking 1 / 2 and 1 / 4 respectively will yield good results.

[0040] Preferably, in step six, the formula for the compensated control quantity is as follows:

[0041]

[0042] The force feedback nonlinear active disturbance rejection controller includes a total of , , , , , , , , , and These 11 sets of parameters.

[0043] Preferably, the optimization method for the particle swarm optimization is as follows:

[0044] S1: Set the constant inertia factor in the algorithm constant of acceleration and The position variables of particles in the search space are generated through random methods. and the speed of ion movement ;

[0045] S2: Parameters in the tracking differentiator and , and parameters in the extended state observer , , and As the optimization target, it corresponds to the position variables of different particles in the particle swarm. After algorithmic calculation, the position variables of each particle in the particle swarm are... Assign the corresponding parameters to the tracking differentiator and the extended state observer, respectively;

[0046] S3: Utilizing parameters optimized by the particle swarm optimization algorithm , , , , and Simulations were performed in a nonlinear active disturbance rejection controller and force feedback system to obtain the control output. ;

[0047] S4: Determine whether the control effect meets the conditions based on the time integral of the absolute error. If it does, the optimization ends; otherwise, the particle population is updated and optimization continues.

[0048] Preferably, the ITAE index is calculated using the following formula:

[0049]

[0050] The formula for particle swarm optimization updates is as follows:

[0051]

[0052] in, This represents the locally optimal position found by the particle at this moment. It is the globally optimal position found throughout the entire search process.

[0053] Compared with the prior art, the beneficial effects of the present invention are as follows: The present invention uses an extended state observer for real-time estimation to compensate for the total disturbances experienced by the force feedback system during operation. In addition, the designed nonlinear active disturbance rejection control technology can perform force feedback control without relying on a specific system model, and has strong adaptability and robustness. The particle swarm optimization algorithm can automatically perform parameter tuning, which improves the design efficiency of the nonlinear active disturbance rejection controller and optimizes the control performance. Attached Figure Description

[0054] Figure 1 This is a structural block diagram of the present invention.

[0055] Figure 2 This is a flowchart of the method of the present invention;

[0056] Figure 3 This is a control flow diagram of the present invention;

[0057] Figure 4 This is a simulation connection diagram of the present invention;

[0058] Figure 5This is a flowchart of the particle swarm optimization algorithm of the present invention;

[0059] Figure 6 This is the dynamic model of the feedback force feedback manipulator control system of the present invention.

[0060] In the diagram: 1. Signal input module; 2. Tracking differentiator; 3. Feedback control law; 4. Extended state observer; 5. Force feedback system; 6. Particle swarm algorithm module. Detailed Implementation

[0061] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0062] Please see Figure 1 The present invention provides a technical solution: a force feedback nonlinear active disturbance rejection controller based on particle swarm optimization, comprising a signal input module 1, a tracking differentiator 2, a feedback control law 3, an extended state observer 4, a force feedback system 5, and a particle swarm algorithm module 6. The output terminals of the signal input module 1 and the particle swarm algorithm module 6 are electrically connected to the tracking differentiator 2. The output terminals of the tracking differentiator 2 and the extended state observer 4 are electrically connected to the feedback control law 3. The output terminal of the feedback control law 3 is electrically connected to the force feedback system 5 and the extended state observer 4. The output terminal of the force feedback system 5 is electrically connected to the particle swarm algorithm module 6 and the extended state observer 4, respectively.

[0063] Please see Figure 2-6 The present invention provides a technical solution: a control method for a force feedback nonlinear active disturbance rejection controller based on particle swarm optimization, comprising the following steps: Step 1, building a model; Step 2, determining the spatial expression; Step 3, transition processing of displacement signals; Step 4, processing of feedback control signals; Step 5, processing of extended state observation signals; and Step 6, compensation of the control quantity feedback system.

[0064] In step one above, the position variables of the generated particles in the search space are obtained through particle swarm optimization. and the velocity of particles A kinematic and dynamic model of the force feedback manipulator was constructed. The kinematic model includes a theoretical model of the spring tension and a displacement model of the motor driven by the track. The formulas are as follows:

[0065]

[0066]

[0067] By combining the two equations and differentiating them separately, we can obtain the speed control quantity for the DC motor:

[0068]

[0069] The dynamic model of the force feedback manipulator is described as follows:

[0070]

[0071] The optimization method for the particle swarm optimization is as follows:

[0072] S1: Set the constant inertia factor in the algorithm constant of acceleration and The position variables of particles in the search space are generated through random methods. and the speed of ion movement ;

[0073] S2: Parameters in tracking differentiator 2 and , and the parameters in extended state observer 4 , , and As the optimization target, it corresponds to the position variables of different particles in the particle swarm. After algorithmic calculation, the position variables of each particle in the particle swarm are... Assign the corresponding parameters to the tracking differentiator 2 and the extended state observer 4 respectively;

[0074] S3: Utilizing parameters optimized by the particle swarm optimization algorithm , , , , and Simulations were performed on a nonlinear active disturbance rejection controller and a force feedback system 5 to obtain the control output. ;

[0075] S4: Determine whether the control effect meets the conditions based on the absolute error time integral (ITAE). If it meets Y, the optimization ends. If it does not meet N, the particle population is updated and the optimization continues.

[0076] The formula for calculating the ITAE metric is as follows:

[0077]

[0078] The formula for particle swarm optimization updates is as follows:

[0079]

[0080] in, This represents the locally optimal position found by the particle at this moment. It is the globally optimal position found throughout the entire search process;

[0081] In step two above, the control output is obtained by optimizing the parameters using the particle swarm optimization method. Determine the state-space expression of the force feedback manipulator and take the state variables. , Therefore, the state equation of force feedback system 5 is:

[0082]

[0083] The state-space expression of force feedback system 5 is as follows:

[0084]

[0085]

[0086] In step three above, signal input module 1 obtains the displacement signal. The input is fed into the tracking differentiator 2, and the output is... The signal is the input displacement signal. The signal after transition The signal is a displacement signal. The differential signal tracks the system state at the output port of differentiator 2. and The discrete formula is as follows:

[0087]

[0088] The formula for tracking differentiator 2 includes a core algorithmic component: the steepest control function. This allows the entire tracking differentiator 2 to quickly and smoothly track changes in the input signal. Its tracking formula is as follows:

[0089]

[0090] The tracking differentiator 2 includes the tracking velocity factor as a parameter. and filter factor ;

[0091] In step four above, after tracking the transition signal of differentiator 2... Signals and The signals are respectively compared with the output of the extended state observer 4 and Subtraction yields the input of feedback control law 3. and Feedback control law 3 is implemented through a nonlinear combination method, and the calculated output is: The expression for feedback control law 3 is as follows:

[0092]

[0093] in It is a nonlinear function:

[0094]

[0095] The parameters included in feedback control law 3 are: , , , and ;

[0096] In step five above, the original state variable is expanded in the extended state observer 4, transforming the total disturbance of the sum of external and internal disturbances into a new state variable. The input signal of the extended state observer 4 is... and The output is , and The following nonlinear extended state observer is established for the force feedback manipulator system:

[0097]

[0098] The discrete form of the extended observer is represented by the following formula:

[0099]

[0100] The parameters included in the extended observer are: , , , , and ,in and It can be set to any value; taking 1 / 2 and 1 / 4 respectively will yield good results.

[0101] In step six above, the compensated control quantity is applied to the force feedback manipulator system to achieve precise control. The formula for the compensated control quantity is as follows:

[0102]

[0103] The force feedback nonlinear active disturbance rejection controller includes a total of , , , , , , , , , and These 11 sets of parameters.

[0104] Based on the above, the advantages of this invention are as follows: The force feedback manipulator control technology developed in this invention is used for precise control of the force feedback manipulator, reducing the interference of disturbance factors on the dynamic model. The active disturbance rejection controller (ADRC) can estimate and compensate for the total disturbance experienced by the force feedback system 5 during operation in real time through the extended state observer 4. Therefore, the ADRC has strong adaptability and robustness, and can perform force feedback control without relying on a specific system model. A particle swarm optimization algorithm is designed for automatic parameter tuning based on nonlinear ADRC, improving the design efficiency of the nonlinear ADRC. The tracking differentiator 2 mainly tracks and smoothly transitions the input displacement signal, thereby preventing oscillations and interference caused by discontinuous or random disturbances in the input control signal of the force feedback system 5. Extracting the continuous differentiable displacement signal can prevent overshoot caused by a large initial control output, thus improving the quality of the force feedback ADRC. The feedback control law 3 is an improvement over the traditional method of using discrete error signals. and discrete error differential signal and the generated discrete error integral signal The improvement to the linear combination PID control, namely, after tracking the transition signal of the differentiator 2, can achieve the feedback control law 3 through nonlinear combination. The core idea of ​​the algorithm of expanding the state observer 4 is to expand the original state variable, transforming the total disturbance of the sum of external disturbance and internal disturbance into a new state variable. The state variable after expanding the state observer 4 includes the original state variable and the state variable of the disturbance observation.

[0105] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the invention can be implemented in other specific forms without departing from its spirit or essential characteristics. Therefore, the embodiments should be considered in all respects as exemplary and non-limiting, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be included within the present invention. No reference numerals in the claims should be construed as limiting the scope of the claims.

Claims

1. A control method for a force feedback nonlinear active disturbance rejection controller based on particle swarm optimization, comprising a signal input module (1), a tracking differentiator (2), a feedback control law (3), an extended state observer (4), a force feedback system (5), and a particle swarm algorithm module (6), characterized in that: The output terminals of the signal input module (1) and the particle swarm algorithm module (6) are electrically connected to the tracking differentiator (2). The output terminals of the tracking differentiator (2) and the extended state observer (4) are electrically connected to the feedback control law (3). The output terminal of the feedback control law (3) is electrically connected to the force feedback system (5) and the extended state observer (4). The output terminal of the force feedback system (5) is electrically connected to the particle swarm algorithm module (6) and the extended state observer (4), respectively. The process includes: Step 1, model building; Step 2, determining the spatial expression; Step 3, transition processing of displacement signals; Step 4, processing of feedback control signals; Step 5, processing of extended state observation signals; and Step 6, compensation of the control quantity feedback system. Its key feature is that: In step one above, the position variables of the generated particles in the search space are obtained through particle swarm optimization. and the velocity of particles Build kinematic and dynamic models of force feedback manipulators; In step two above, the control output is obtained by optimizing the parameters using the particle swarm optimization method. Determine the state-space expression of the force feedback manipulator; In step three above, the signal input module (1) obtains the displacement signal. The input is fed into the tracking differentiator (2), and the output is... The signal is the input displacement signal. The signal after transition The signal is a displacement signal. The differential signal; In step four above, after tracking the transition signal of the differentiator (2), Signals and The signals are respectively compared with the output of the extended state observer (4). and Subtraction yields the input of the feedback control law (3). and The feedback control law (3) is implemented through a nonlinear combination method, and the calculated output is: ; In step five above, the original state variable is expanded in the extended state observer (4), transforming the total disturbance of the sum of external and internal disturbances into a new state variable. The input signal of the extended state observer (4) is... and The output is , and ; In step six above, the compensated control quantity is applied to the force feedback manipulator system to achieve precise control. In step one, the kinematic model includes a theoretical model of the spring tension and a displacement model of the motor driven by the track, with the following formulas: ， ， By combining the two equations and differentiating them separately, we can obtain the speed control quantity for the DC motor: ， The dynamic model of the force feedback manipulator is described as follows. 。 2. The control method for a force feedback nonlinear active disturbance rejection controller based on particle swarm optimization according to claim 1, characterized in that: In step two, the state variable is taken. , Then the state equation of the force feedback system (5) is: ， The state-space expression of the force feedback system (5) is as follows: ， 。 3. The control method for a force feedback nonlinear active disturbance rejection controller based on particle swarm optimization according to claim 1, characterized in that: In step three, the system state of the output port of the differentiator (2) is tracked. and The discrete formula is as follows: ， The formula for the tracking differentiator (2) contains a core algorithmic component, namely the steepest control function. This allows the entire tracking differentiator (2) to quickly and smoothly track changes in the input signal. Its tracking formula is as follows: ， The tracking differentiator (2) includes the tracking velocity factor as a parameter. and filter factor .

4. The control method of the force feedback nonlinear active disturbance rejection controller based on particle swarm optimization according to claim 1, characterized in that: In step four, the expression for the feedback control law (3) is as follows: ， in It is a nonlinear function: ， The parameters included in the feedback control law (3) are: , , , and .

5. The control method for a force feedback nonlinear active disturbance rejection controller based on particle swarm optimization according to claim 1, characterized in that: In step five, a nonlinear extended state observer (4) is established for the force feedback manipulator system as follows: ， The discrete form of the extended observer is represented by the following formula: ， The parameters included in the extended observer are: , , , , and .

6. The control method for a force feedback nonlinear active disturbance rejection controller based on particle swarm optimization according to claim 1, characterized in that: In step six, the formula for the compensated control quantity is as follows: ， The force feedback nonlinear active disturbance rejection controller includes a total of , , , , , , , , , and These 11 sets of parameters.

7. The control method for a force feedback nonlinear active disturbance rejection controller based on particle swarm optimization according to claim 1, characterized in that: The optimization method for the particle swarm optimization is as follows: S1: Set the constant inertia factor in the algorithm constant of acceleration and The position variables of particles in the search space are generated through random methods. and the speed of ion movement ; S2: Parameters in the tracking differentiator (2) and , and the parameters in the extended state observer (4) , , and As the optimization target, it corresponds to the position variables of different particles in the particle swarm. After algorithmic calculation, the position variables of each particle in the particle swarm are... Assign the corresponding parameters to the tracking differentiator (2) and the extended state observer (4), respectively; S3: Utilizing parameters optimized by the particle swarm optimization algorithm , , , , and Simulations were performed on the nonlinear active disturbance rejection controller and force feedback system (5) to obtain the control output. ; S4: Determine whether the control effect meets the conditions based on the time integral of absolute error (ITAE). If it meets the conditions (Y), the optimization ends. If it does not meet the conditions (N), the particle population is updated and the optimization continues.

8. The control method for a force feedback nonlinear active disturbance rejection controller based on particle swarm optimization according to claim 7, characterized in that: The formula for calculating the absolute error time integral index is as follows: ， The formula for particle swarm optimization updates is as follows: ， in, This represents the locally optimal position found by the particle at this moment. It is the globally optimal position found throughout the entire search process.

Images