Self-adaptive control method for attitude of servo robot arm holder

Abstract

The invention discloses a servo manipulator holder attitude adaptive control method, and belongs to the field of holder attitude control, and the method comprises the steps: S1, inputting a three-dimensional coordinate point (x1, y1, z1) of a target position of a servo manipulator, and enabling a coordinate point x1 of a horizontal axis and a coordinate point y1 of a longitudinal axis of the three-dimensional coordinate point to form a two-dimensional coordinate point of the target position; s2, according to the two-dimensional coordinate point and a two-dimensional coordinate point (x2, y2) of the current position of the servo manipulator, an xy plane attitude deflection angle of the target position and the current position of the servo manipulator is calculated; s3, inputting the xy plane attitude deflection angle into the pan-tilt attitude conversion mathematical model, and outputting a pan-tilt rotation angle error value; s4, inputting the rotation angle error value of the holder into an improved PID (Proportion Integration Differentiation) algorithm; and S5, applying the improved PID algorithm to a pan-tilt attitude control system, and outputting a pan-tilt driving value to adjust the pan-tilt rotation angle according to the real-time pan-tilt rotation angle error value so as to realize adaptive accurate and stable control of the pan-tilt attitude.

Description

Technical Field

[0001] The invention relates to the field of pan-tilt posture control, and in particular to a pan-tilt posture adaptive control method for a servo robot arm. Background Art

[0002] In actual work, the servo manipulator mainly relies on the chassis pan-tilt adjustment to accurately reach the target position. When facing a complex and changeable environment, the traditional pan-tilt control system cannot make posture adjustments quickly and accurately, resulting in poor grasping robustness of the servo manipulator; secondly, in a complex working environment, external interference (such as mechanical vibration, electromagnetic interference) often causes the pan-tilt posture to shift. The traditional pan-tilt control system has limited filtering and compensation capabilities for these external interferences, and is prone to control lag or insufficient posture adjustment, which in turn affects the accuracy of the servo manipulator; furthermore, most traditional control methods rely on preset control models, such as the standard PID control algorithm, which lacks sufficient real-time feedback mechanism, and the proportional parameter Kp, integral parameter Ki and differential parameter Kd are fixed, and real-time adaptive adjustment in a dynamic environment cannot be achieved. The manipulator needs to continuously correct its motion trajectory and posture based on sensor feedback data, and the traditional pan-tilt control system is relatively slow in this regard. When the manipulator load changes, the steady-state error of the pan-tilt posture control is also large.

[0003] The wave search algorithm (WSA) is an optimization method inspired by radar technology, which aims to solve complex optimization problems. The algorithm combines the advantages of deterministic optimization technology and uncertain optimization technology, and realizes effective exploration and utilization of the search space by simulating the emission, reflection and reception process of radar wave particle swarm. The WSA algorithm adopts a unique initialization method to generate the initial particle swarm, which helps to increase the diversity of the particle swarm. It combines the gradient descent method of fitting and uses the analytical information of the problem to improve the search efficiency and accuracy. The problem with WSA is that the parameters in the algorithm (such as initial step size, step size scaling factor) have a great influence on the algorithm performance, and tedious adjustments are required to obtain the best algorithm optimization performance. Improper parameter settings will cause the algorithm to converge more slowly or fall into a local optimal solution. At the same time, although the algorithm adopts a generalized boundary restriction rule to deal with out-of-bounds particles, this strategy reduces the diversity of the wave particle swarm to a certain extent and affects the search performance. Summary of the invention

[0004] The present invention aims at the problem of poor control robustness of a traditional manipulator pan-tilt when the manipulator load changes and the problem of PID control performance defects of a standard wave search algorithm (WSA), and provides a servo manipulator pan-tilt attitude adaptive control method. By improving the wave search algorithm (WSA), an improved wave search algorithm (GWSA) is used to optimize the control parameters of a position-based PID algorithm to obtain an enhanced PID algorithm, wherein the control parameters include: a proportional parameter Kp, an integral parameter Ki and a differential parameter Kd. The pan-tilt system of the servo manipulator is controlled by the enhanced PID algorithm to achieve precise control of the pan-tilt attitude of the servo manipulator.

[0005] In order to achieve the above-mentioned purpose, the technical solution adopted by the present invention is as follows: a servo robot arm pan-tilt attitude adaptive control method, using the excellent optimization performance of the improved wave search algorithm (GWSA) to find the optimal Kp, Ki, and Kd control parameters of the PID algorithm of the pan-tilt attitude control system of the servo robot arm, and the specific steps are as follows.

[0006] S1. Input the three-dimensional coordinate point (x1, y1, z1) of the target position of the servo robot arm, and combine the horizontal axis coordinate point x1 and the vertical axis coordinate point y1 of the three-dimensional coordinate point to form the two-dimensional coordinate point (x1, y1) of the target position.

[0007] S2. Calculate the xy plane attitude deviation angle between the target position and the current position of the servo robot arm according to the two-dimensional coordinate point (x1, y1) of the target position and the two-dimensional coordinate point (x2, y2) of the current position of the servo robot arm.

[0008] S3, the xy plane attitude deflection angle is input into the gimbal attitude conversion mathematical model, and the gimbal rotation angle error value is output.

[0009] S4. The pan / tilt rotation angle error value is input into an improved PID algorithm. The improved PID algorithm is specifically as follows: the Kp, Ki, and Kd control parameters of the position PID are adjusted using an improved wave search algorithm to obtain optimal control parameters, and the optimal control parameters are used in the position PID algorithm to obtain an improved PID algorithm; the improved wave search algorithm includes: S41, establishing a heuristic guidance strategy, specifically: guiding the wave search algorithm to optimize the position PID control parameters by introducing heuristic rules, and improving the optimization mechanism of the wave search algorithm; S42. Improve the waveform size control coefficient α based on the dynamic sinusoidal strategy of the variance of fitness values ​​between wave particles and the global minimum fitness value of wave particles, and use the improved waveform size control coefficient to improve the wave particle position update mathematical model in the local development stage of the wave search algorithm.

[0010] S5. Use the improved PID algorithm for the gimbal attitude control system and output the gimbal drive value according to the real-time gimbal rotation angle error value. Adjust the gimbal rotation angle to achieve adaptive, precise and stable control of the gimbal attitude.

[0011] Preferably, the pan-tilt attitude control system comprises a pan-tilt of a servo robot and a pan-tilt drive unit, wherein the pan-tilt of the servo robot is a brushless DC motor, and the pan-tilt drive unit comprises an enhanced PID algorithm and a pan-tilt attitude conversion mathematical model; The gimbal attitude conversion mathematical model converts the xy plane attitude deflection angle into a gimbal rotation angle error value, and the gimbal attitude control system adjusts the gimbal attitude to reach the target attitude according to the gimbal rotation angle error value; wherein, the enhanced PID algorithm of the gimbal drive unit is responsible for calculating the gimbal drive signal value of the servo robot arm.

[0012] Preferably, the enhanced PID algorithm mathematical model is: (8); In formula (8), For the The gimbal drive value in seconds, is the running time, For the The gimbal rotation angle error value in seconds, is the optimal ratio parameter, is the optimal integration parameter, is the optimal differentiation parameter.

[0013] Preferably, the xy plane attitude deflection angle of the servo manipulator is converted into a pan-tilt rotation angle error value through a pan-tilt attitude conversion mathematical model, wherein the pan-tilt attitude conversion mathematical model is: (9); In formula (9), For the The gimbal rotation angle error value in seconds, , , are the attitude error weight, attitude error change rate weight and acceleration weight, For the Seconds of xy plane attitude deflection, is the rate of change of the xy plane attitude error, is the acceleration of the attitude error in the xy plane.

[0014] Preferably, the wave search algorithm is improved, wherein a heuristic guidance strategy is introduced, with the first purpose of increasing the diversity of the wave particle group of the wave search algorithm, thereby providing more optimal control parameter candidate solutions for the improved wave search algorithm to optimize the position PID control parameters; the second purpose is to accelerate the local convergence speed of the wave search algorithm, thereby improving the accuracy of Kp, Ki, and Kd parameters of the position PID obtained by the improved wave search algorithm to optimize the position PID control parameters, and further, the waveform size control coefficient α is improved, with the first purpose of balancing the global search and local convergence stages of the wave search algorithm, thereby avoiding the improved wave search algorithm from falling into the local optimal value during the optimization process, improving the accuracy of Kp, Ki, and Kd parameters of the position PID, and finally achieving precise and stable control of the gimbal's attitude.

[0015] Preferably, a heuristic guidance strategy is designed to guide the search process of the wave search algorithm by introducing heuristic rules to improve the optimization convergence speed of the wave search algorithm. The guiding function h used by the heuristic rule reduces the search space of the wave search algorithm according to the random position of the wave particles. The mathematical model is: (1); In formula (1), is the j-dimensional position value of the randomly generated i-th wave particle individual, that is, the random candidate solution value, A heuristic bootstrap function that adjusts the value of a random candidate solution, where To guide the strength coefficient and control the adjustment range, is a heuristic information function, which is used to represent the adjustment amount from the random candidate solution value to the center of the local area range R.

[0016] Preferably, the local area range R is represented by a Gaussian distribution, and the center of the local area range R, i.e., the individual optimal position of the wave search algorithm at the tth iteration, is expressed as , the covariance matrix is ​​Σ, and the mathematical model is established to define the heuristic information function as: (2); In formula (2), is the individual best position of the wave search algorithm at the tth iteration, is the transpose operation.

[0017] Preferably, the heuristic guidance strategy is used in the initialization phase and the local development phase of the wave search algorithm, and the specific method is: S311, when initializing the population position of the wave search algorithm in each iteration, the guide function h(x) is used to generate a candidate solution set in the search space of the wave search algorithm, and the mathematical model is: (3); In formula (3), is the position of the i-th wave particle individual in the candidate set, is the j-dimensional position value of the randomly generated i-th wave particle individual, that is, the random candidate solution value, A heuristic bootstrap function that adjusts the values ​​of random candidate solutions; S312. In the local development stage, the guide function h(x) is used to update the candidate solution set, which is the wave particle population position. The mathematical model is: (4); In formula (4), is the position value of the j-th dimension of the ith wave particle in the t+1th iteration, is the position value of the j-th dimension of the ith wave particle in the t-th iteration, is the maximum number of iterations, For A heuristic guidance function to adjust the solution value; is the global optimal position of individual wave particles.

[0018] Preferably, the waveform size control coefficient α is improved based on the dynamic sinusoidal strategy of the variance of the fitness values ​​between wave particles and the global minimum fitness value of the wave particles. The mathematical model of the improved waveform size control coefficient α(t) is: (5); In formula (5), is the initial value of the improved waveform size control coefficient, is the variance of the fitness value, is the average fitness value, is the fitness value, is the global minimum fitness value of the wave particle, is the minimum fitness value of the tth iteration, where The mathematical model is: (6); In formula (6), N is the size of the wave particle population, is the fitness value of the i-th wave particle position.

[0019] Preferably, dynamic adjustment based on fitness variance enables the improved wave search algorithm to adaptively adjust the waveform size control coefficient α(t) according to the diversity of the wave particle population. When the wave particle population shows high diversity, α(t) increases, thereby increasing the improved wave search algorithm's search range and ability to explore different areas in the early stage of iteration; at the same time, the waveform size control coefficient α(t) can be dynamically adjusted according to the convergence speed of the improved wave search algorithm, thereby refining the search range in the later stage of the search and improving the accuracy of the solution. When the algorithm quickly approaches the optimal solution, reducing α(t) can reduce overshoot and oscillation in the optimization process of the improved wave search algorithm, helping the improved wave search algorithm to converge stably to the global optimal solution; finally, introducing random perturbation elements helps the algorithm maintain a certain degree of randomness and flexibility when facing complex search spaces and multiple local optima. Random perturbations enable the improved wave search algorithm to jump out and continue to explore other optimal solution areas when converging to a local optimal solution.

[0020] Preferably, the wave particle position update mathematical model of the local development stage of the wave search algorithm is improved by using the improved waveform size control coefficient, specifically the mathematical model of the transmitted radar wave in the local development stage; (7); In formula (7), is the new position of the ith wave particle at the tth iteration, is the current optimal wave particle position, is the current second-best wave particle position, is a column vector whose elements follow a normal distribution. is the improved waveform size control coefficient value of the tth iteration.

[0021] Preferably, the Kp, Ki, and Kd parameters of the position PID are adjusted using an improved wave search algorithm to obtain the optimal control parameters, wherein the Kp, Ki, and Kd parameters of the position PID are mapped as a space vector to the wave particle position of the improved wave search algorithm, and the specific steps are as follows: P1. Initialize the parameters of the improved wave search algorithm, including: wave particle population size N, problem dimension D, search space upper bound UB 、 The search space lower bound LB and the maximum number of iterations T max and improved waveform size control coefficients; P2, encoding the Kp, Ki, Kd proportional coefficients of the position PID algorithm into N groups of three-dimensional space vectors, and the space vectors are mapped to the real numbers of the wave particle positions; P3, randomly initialize the initial position of the improved wave search algorithm population, and use the mathematical model of formula (3) to generate candidate solutions using the guide function h(x) in the search space of the wave search algorithm , where the upper and lower bounds of the search space are the value ranges of the three-dimensional space vectors; P4. Determine whether the current number of iterations t is equal to the maximum number of iterations T max If yes, then output the wave particle position corresponding to the minimum fitness value and resolve it into the optimal Kp, Ki, Kd proportional coefficient values ​​of the position PID algorithm; otherwise, return to execute P5; P5. Establish an improved wave search algorithm position update mathematical model to update the wave particle position, including: P51, by introducing the random number r1 and the current number of iterations, a mathematical model for position update in the global search phase is established to generate the new position of the wave particle; P52, simulate the radar wave emission process, establish a mathematical model according to formula (7) to generate new wave particle positions, and explore the potential optimal solution area; P53, simulate the reflection and reception process of radar waves, establish a mathematical model for the local development stage to update the position of the wave particle population, and gradually approach the global optimal solution by updating the wave particle position through multiple iterations. The global optimal solution is the wave particle position with the smallest fitness value in the wave particle population; P54, using the mathematical model formula (4) to further update the current wave particle population position; P6. Use the objective function to calculate the fitness value of each wave particle position, retain the wave particle position corresponding to the minimum fitness value from initialization to the tth iteration, execute t=t+1 for the current iteration number t, and return to execute step P3.

[0022] Preferably, the stability of the gimbal during the posture adjustment process is ensured to avoid sudden movement and shaking, while ensuring that the gimbal can accurately reach the target posture angle, and the objective function J is designed, and its mathematical model is: (10); In formula (10), is the running time, For the The actual gimbal attitude angle in seconds, For the PTZ target attitude angle in seconds, It is the average value of the actual attitude angle of the gimbal.

[0023] Preferably, the fitness value is an indicator for evaluating the control effect of each group of PID control parameters. The fitness value is calculated by the objective function, specifically: each wave particle position is parsed into a group of three-dimensional space vectors, that is, each space vector includes Kp, Ki, Kd values, and the Kp, Ki, Kd values ​​are input into the position PID control algorithm for the entire gimbal attitude control system, and the objective function value is calculated according to the real-time attitude response output by the attitude control system, which is recorded as the fitness function value. The attitude response is the first The actual attitude angle in seconds.

[0024] Preferably, the improved wave search algorithm (GWSA) is mainly aimed at improving the global exploration information of the algorithm and the diversity of the wave particle group in the global search stage. By introducing random numbers and the current number of iterations, a mathematical model for position update in the global search stage is established. The mathematical model is: (11); In formula (11), is the new position of the ith wave particle at the tth iteration, is the minimum position of the ith wave particle, is the maximum position of the i-th wave particle, A random number between 0 and 1.

[0025] Preferably, the improved wave search algorithm (GWSA) is divided into the reflection of radar waves and the reception of radar waves in the local development stage, wherein the reflection simulation part of radar waves establishes a wave-particle position mathematical model according to formula (12) when particles encounter obstacles and reflect back to the optimal position; (12); In formula (12), is the new position of the ith wave particle at the tth iteration, β is the reflection intensity factor, is a random number between 0 and 1. is the total number of wave particles simulating the reflected electromagnetic wave, is the current optimal wave particle position, is the position matrix rearranged from minimum to maximum.

[0026] Preferably, according to formula (13), the radar wave receiving process simulates the radar receiving electromagnetic waves to establish a wave-particle position mathematical model, and the wave-particle position is corrected and processed; (13); In formula (12), is the new position of the ith wave particle at the tth iteration, is the position of the ith wave particle before update in the tth iteration, is a random number that follows a normal distribution, is the receiving strength factor, is the current optimal wave particle position, To simulate the total number of wave particles received by the electromagnetic wave, , , and is a random number between 0 and 1. is the receiving weight factor.

[0027] In summary, due to the adoption of the above technical scheme, the beneficial effects of the present invention are as follows: the present invention improves the optimization accuracy and speed of the standard wave search algorithm at different stages through the improved waveform size control coefficient α and the introduced heuristic guidance strategy, and uses the improved wave search algorithm to optimize the Kp, Ki, Kd control parameters of the position PID to obtain an improved PID algorithm, which improves the response speed and accuracy of the gimbal attitude control system of the servo robot arm, and can accurately control the attitude angle of the gimbal, so that when the load of the robot arm changes, the gimbal control robustness reaches the optimal level. At the same time, the improved PID algorithm can better adapt to environmental changes, and the attitude control of the gimbal can remain stable in a complex working environment. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] Figure 1 Flowchart of the gimbal posture control optimization method for a servo robot arm.

[0029] Figure 2 Graph showing the change in fitness value of wave particle position during the optimization process of the improved wave search algorithm and the standard wave search algorithm.

[0030] Figure 3 This is a graph showing the changes in PID control parameters during the optimization process of the standard wave search algorithm.

[0031] Figure 4 The diagram showing the changes in PID control parameters during the optimization process for the improved wave search algorithm.

[0032] Figure 5 This is the effect diagram of the servo robot arm's pan-tilt attitude adaptive control. DETAILED DESCRIPTION

[0033] 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.

[0034] See also Figure 1-Figure 5The present invention provides a technical solution: a servo robot arm pan-tilt attitude adaptive control method, establishes a servo robot arm pan-tilt attitude control system, the control system includes a servo robot arm pan-tilt and a pan-tilt drive unit, wherein the servo robot arm pan-tilt is a brushless DC motor, the pan-tilt drive unit includes an enhanced PID algorithm and a pan-tilt attitude conversion mathematical model, and uses the optimization performance of an improved wave search algorithm (GWSA) to find the optimal Kp, Ki, and Kd control parameters of the PID control algorithm of the servo robot arm pan-tilt attitude control system. The specific steps are as follows: Figure 1 shown.

[0035] Furthermore, in step S1, the three-dimensional coordinate point (x1, y1, z1) of the target position of the servo robot arm is input, and the horizontal axis coordinate point x1 and the vertical axis coordinate point y1 of the three-dimensional coordinate point are combined into the two-dimensional coordinate point (x1, y1) of the target position.

[0036] Furthermore, in step S2, the xy plane posture deviation angle between the target position and the current position of the servo robot is calculated according to the two-dimensional coordinate point (x1, y1) of the target position and the two-dimensional coordinate point (x2, y2) of the current position of the servo robot.

[0037] More specifically, the xy plane attitude deflection is expressed in the range of 0 to 360 degrees, using the formula Calculate the Xy plane attitude deflection angle in seconds .

[0038] Furthermore, in step S3, the xy plane attitude deflection angle is input into the gimbal attitude conversion mathematical model, and the gimbal rotation angle error value is output.

[0039] More specifically, the mathematical model of gimbal attitude conversion is used to convert the first Xy plane attitude deflection angle in seconds Convert to gimbal rotation angle error value; (9); In formula (9), For the The gimbal rotation angle error value in seconds, , , are the attitude deflection weight, attitude error change rate weight and acceleration weight, For the Seconds of xy plane attitude deflection, is the rate of change of the attitude deflection angle in the xy plane, is the acceleration of the attitude deflection in the xy plane.

[0040] More specifically, in Matlab, 10 sets of experimental data of gimbal rotation angle error and xy plane attitude deflection are set, and the change rate and acceleration of each set of xy plane attitude deflection are calculated by derivation, and the xy plane attitude deflection in each set of experimental data, as well as the change rate of the xy plane attitude deflection and the acceleration of the xy plane attitude deflection are converted into matrices , the gimbal rotation angle error value corresponding to each set of experimental data is taken as the target response , use the least squares method to estimate the parameters , , , the mathematical model is: , where T is the transposition operation. The estimated model parameters are implemented into the gimbal attitude conversion mathematical model of the gimbal attitude control system to obtain the attitude error weight =0.00001, weight of attitude error change rate = 1.4 and the weight of the attitude error acceleration =2.7.

[0041] Further, in step S4, the pan / tilt rotation angle error value is input into the improved PID algorithm, wherein the improved PID algorithm is specifically: the Kp, Ki, and Kd control parameters of the position PID are adjusted by using the improved wave search algorithm to obtain the optimal control parameters, and the optimal control parameters are used in the position PID algorithm to obtain the improved PID algorithm; the improved wave search algorithm includes: S41, establishing a heuristic guidance strategy, specifically: guiding the wave search algorithm to optimize the position PID control parameters by introducing heuristic rules, and improving the optimization mechanism of the wave search algorithm; S42. Improve the waveform size control coefficient α based on the dynamic sinusoidal strategy of the variance of fitness values ​​between wave particles and the global minimum fitness value of wave particles, and use the improved waveform size control coefficient to improve the wave particle position update mathematical model in the local development stage of the wave search algorithm.

[0042] More specifically, a heuristic guidance strategy is designed to guide the search process of the wave search algorithm by introducing heuristic rules to improve the optimization convergence speed of the wave search algorithm. The guiding function h used by the heuristic rule narrows the search space of the wave search algorithm according to the random position of the wave particles. The mathematical model is: (1); In formula (1), is the j-dimensional position value of the randomly generated i-th wave particle individual, that is, the random candidate solution value, A heuristic bootstrap function that adjusts the value of a random candidate solution, where To guide the strength coefficient and control the adjustment range, is a heuristic information function, which is used to represent the center adjustment amount from the random candidate solution value to the local area range R.

[0043] More specifically, the local region range R is represented by a Gaussian distribution, and the center of the local region range R, i.e., the individual optimal position of the wave search algorithm at the tth iteration, is expressed as , the covariance matrix is ​​Σ, and the mathematical model is established to define the heuristic information function as: (2); In formula (2), is the heuristic information function, is the individual best position of the wave search algorithm at the tth iteration, is the transpose operation.

[0044] More specifically, the heuristic guidance strategy is used in the initialization phase and the local development phase of the wave search algorithm, and the specific method is as follows: S311, when initializing the population position of the wave search algorithm in each iteration, the guide function h(x) is used to generate a candidate solution set in the search space of the wave search algorithm, and the mathematical model is: (3); In formula (3), is the position of the i-th wave particle individual in the candidate set, is the j-dimensional position value of the randomly generated i-th wave particle individual, that is, the random candidate solution value, A heuristic bootstrap function that adjusts the values ​​of random candidate solutions; S312. In the local development stage, the guide function h(x) is used to update the candidate solution set, which is the wave particle population position. The mathematical model is: (4); In formula (4), is the position value of the j-th dimension of the ith wave particle in the t+1th iteration, is the position value of the j-th dimension of the ith wave particle in the t-th iteration, is the maximum number of iterations, For A heuristic guidance function to adjust the solution value; is the global optimal position of individual wave particles.

[0045] More specifically, the waveform size control coefficient α is improved based on the dynamic sinusoidal strategy of the variance of the fitness values ​​between wave particles and the global minimum fitness value of the wave particles. The mathematical model of the improved waveform size control coefficient α(t) is: (5); In formula (5), is the initial value of the improved waveform size control coefficient, is the variance of the fitness value, is the average fitness value, is the fitness value, is the global minimum fitness value of the wave particle, is the minimum fitness value of the tth iteration, where The mathematical model is: (6); In formula (6), N is the size of the wave particle population, is the fitness value of the i-th wave particle position.

[0046] More specifically, the wave particle position update mathematical model of the local development phase of the wave search algorithm is improved by using the improved waveform size control coefficient, specifically the transmitted radar wave mathematical model of the local development phase; (7); In formula (7), is the new position of the ith wave particle at the tth iteration, is the current optimal wave particle position, is the current second-best wave particle position, is a column vector whose elements follow a normal distribution. is the improved waveform size control coefficient value of the tth iteration.

[0047] More specifically, the improved wave search algorithm is used to adjust the Kp, Ki, and Kd parameters of the position PID to obtain the optimal control parameters, wherein the Kp, Ki, and Kd parameters of the position PID are mapped as a space vector to the wave particle position of the improved wave search algorithm. The specific steps are: P1. Initialize the parameters of the improved wave search algorithm, including: wave particle population size N, problem dimension D, search space upper bound UB 、 The search space lower bound LB and the maximum number of iterations T max and improved waveform size control coefficients; P2, encoding the Kp, Ki, Kd proportional coefficients of the position PID algorithm into N groups of three-dimensional space vectors, and the space vectors are mapped to the real numbers of the wave particle positions; P3, randomly initialize the initial position of the improved wave search algorithm population, and use the mathematical model of formula (3) to generate candidate solutions using the guide function h(x) in the search space of the wave search algorithm , where the upper and lower bounds of the search space are the value ranges of the three-dimensional space vectors; P4. Determine whether the current number of iterations t is equal to the maximum number of iterations T maxIf yes, then output the wave particle position corresponding to the minimum fitness value and resolve it into the Kp, Ki, Kd proportional coefficient values ​​of the position PID algorithm; otherwise, return to execute P5; P5. Establish an improved wave search algorithm position update mathematical model to update the wave particle position, including: P51, by introducing the random number r1 and the current number of iterations, a mathematical model for position update in the global search phase is established to generate a new position of the wave particle; the mathematical model for position update is: (11); In formula (11), is the new position of the ith wave particle at the tth iteration, is the minimum position of the ith wave particle, is the maximum position of the i-th wave particle, A random number between 0 and 1; P52, simulate the radar wave emission process, establish a mathematical model according to formula (7) to generate new wave particle positions, and explore the potential optimal solution area; P53, simulate the reflection and reception process of radar waves, establish a mathematical model of the local convergence stage to update the position of the wave particle population, and gradually approach the global optimal solution by updating the wave particle position through multiple iterations. The global optimal solution is the wave particle position with the smallest fitness value in the wave particle population; wherein, the reflection of radar waves is established according to formula (12) to establish a mathematical model of the wave particle position when the particle encounters an obstacle and reflects back to the optimal position; (12); In formula (12), is the new position of the ith wave particle at the tth iteration, β is the reflection intensity factor, is a random number between 0 and 1. is the total number of wave particles simulating the reflected electromagnetic wave, is the current optimal wave particle position, is the position matrix rearranged from minimum to maximum; In the radar wave receiving stage, the radar receives electromagnetic waves according to formula (13) to establish a wave-particle position mathematical model, and correct and process the wave-particle position; (13); In formula (13), is the new position of the ith wave particle at the tth iteration, is the position of the ith wave particle before update in the tth iteration, is a random number that follows a normal distribution, is the receiving strength factor, is the current optimal wave particle position, To simulate the total number of wave particles received by the electromagnetic wave, , , and is a random number between 0 and 1. is the receiving weight factor; P54, using the mathematical model formula (4) to further update the current wave particle population position; P6. Calculate the fitness value of each wave particle position using the objective function J, retain the wave particle position corresponding to the minimum fitness value from initialization to the tth iteration, execute t=t+1 for the current iteration number t, and return to execute step P3; wherein, the objective function J, its mathematical model is: (10); In formula (10), is the running time, For the The actual attitude angle in seconds, For the Target attitude angle in seconds, is the average value of the actual attitude angle.

[0048] Further, in step S5, the improved PID algorithm is used for the gimbal attitude control system, and the gimbal drive value is output according to the real-time gimbal rotation angle error value. Adjust the gimbal rotation angle to achieve adaptive, precise and stable control of the gimbal attitude.

[0049] Furthermore, by calculating the xy plane attitude deflection angle between the two-dimensional coordinate point (x2, y2) of the current position of the servo robot and the two-dimensional coordinate point (x1, y1) of the target position in real time, the gimbal rotation angle error value is updated in real time through the gimbal attitude conversion mathematical model, and the gimbal drive value is output using formula (8). The mathematical model is: (8); In formula (8), For the The gimbal drive value in seconds, is the running time, For the The gimbal rotation angle error value in seconds, is the optimal ratio parameter, is the optimal integration parameter, is the optimal differentiation parameter.

[0050] Furthermore, in order to verify the advancedness of the method of the present invention, the standard wave search algorithm (WSA) was tested in Matlab and Simulink during the implementation process. First, the population size of the three algorithms was set to N = 100, and the maximum number of iterations Tmax =60, problem dimension D=3, search space upper bound UB=[150,150,150] 、 The search space lower bound LB=[0,0,0], the maximum significant wave height is 2 and the minimum is 0, and three algorithm codes are written; a system simulation model is built in Simulink, and the fitness values ​​of the position PID parameters Kp, Ki, and Kd of the gimbal attitude control system are adjusted using the objective function calculation algorithm. Among them, the standard wave search algorithm (WSA) adjusts the position PID parameters of the gimbal attitude control system. The results are as follows Figure 3 As shown, Kp=71.0795; Ki=0.299346; Kd=0.849411; the results of adjusting the position PID parameters of the gimbal attitude control system by the improved wave search algorithm (GWSA) are as follows Figure 4 As shown, Kp=7.94034; Ki=0.71341; Kd=4.35296.

[0051] Furthermore, if Figure 2 As shown in the figure, the fitness value change diagram of the wave particle position during the optimization process of the improved wave search algorithm and the standard wave search algorithm, the fitness value of the wave search algorithm (WSA) gradually decreases, but the decrease is small, the improved wave search algorithm (GWSA) rapidly decreases faster in the early stage, which indicates that GWSA can find a better solution in the initial stage; between the 10th and 20th iterations, the fitness value of GWSA decreases slowly, showing a stable state, and after about the 30th iteration, the fitness value of GWSA tends to be stable. In contrast, the fitness value of WSA decreases less during the whole process and finally converges to about 15; in terms of convergence speed, GWSA is obviously better than WSA, GWSA reaches a lower fitness value in fewer iterations, showing better optimization ability; secondly, the final fitness value of GWSA is much lower than that of WSA, which means that when the algorithm optimizes the control parameters of the position PID control algorithm, GWSA finds a better parameter combination, thereby achieving more precise gimbal attitude control.

[0052] Furthermore, if Figure 5As shown in the figure, the response curves of three different PID algorithms in gimbal attitude control, the BWO-PID control method has the highest overshoot, and the overshoot gimbal attitude angle value reaches nearly 5 degrees, which indicates that the BWO-PID has a large overshoot phenomenon in the initial stage, resulting in system instability. After a certain period of time, the attitude angle value gradually decreases and tends to stabilize, but the stabilization time is long. The GWSA-PID control method has the smallest overshoot and reaches the target attitude angle value of 4 degrees in the fastest time, which indicates that the GWSA-PID can better suppress overshoot during the control process and ensure system stability. The overshoot of the WSA-PID control method is relatively small, with a peak value close to 4.7, but it is still higher than the GWSA-PID. It enters the stable state slightly faster than the BWO-PID in terms of stabilization time, but the stable angle value is higher than the GWSA-PID. In summary, the GWSA-PID is superior to the BWO-PID and WSA-PID in all aspects. The GWSA-PID shows a faster response speed, the smallest overshoot and steady-state error, and can quickly stabilize at the target attitude angle, which indicates that the GWSA-PID has better performance in attitude control and is more suitable for gimbal attitude control systems that require high precision and stability.

Claims

1. A servo robot arm gimbal attitude adaptive control method, characterized in that: The specific steps are: S1, input the three-dimensional coordinate point (x1, y1, z1) of the target position of the servo robot, where z1 is the z-axis coordinate point of the three-dimensional coordinate point, and combine the horizontal axis coordinate point x1 and the vertical axis coordinate point y1 of the three-dimensional coordinate point to form the two-dimensional coordinate point (x1, y1) of the target position; S2, according to the two-dimensional coordinate point (x1, y1) of the target position and the two-dimensional coordinate point (x2, y2) of the current position of the servo robot, calculate the xy plane posture deviation angle between the target position and the current position of the servo robot; S3, the xy plane attitude deflection angle is input into the gimbal attitude conversion mathematical model, and the gimbal rotation angle error value is output; S4. The pan-tilt rotation angle error value is input into the improved PID algorithm. The improved PID algorithm is specifically: using the improved wave search algorithm to adjust the Kp, Ki, Kd control parameters of the position PID to obtain the optimal control parameters, and using the optimal control parameters for the position PID algorithm to obtain the improved PID algorithm; the improved wave search algorithm includes: S41. Establishing a heuristic guidance strategy, specifically: guiding the wave search algorithm to optimize the position PID control parameters by introducing heuristic rules, and improving the optimization mechanism of the wave search algorithm; S42. Improving the waveform size control coefficient α based on the dynamic sinusoidal strategy of the variance of the fitness values ​​between wave particles and the global minimum fitness value of the wave particles, and using the improved waveform size control coefficient to improve the wave particle position update mathematical model in the local development stage of the wave search algorithm; S5. Use the improved PID algorithm for the gimbal attitude control system and output the gimbal drive value according to the real-time gimbal rotation angle error value. Adjust the gimbal rotation angle to achieve adaptive, precise and stable control of the gimbal attitude.

2. The method for adaptively controlling the posture of a servo robot hand platform according to claim 1, characterized in that: The wave search algorithm is improved in S4, wherein S41 establishes a heuristic guidance strategy, which guides the search process of the wave search algorithm by introducing heuristic rules to improve the optimization convergence speed of the wave search algorithm. The guiding function h used by the heuristic rule reduces the search space of the wave search algorithm according to the random position of the wave particles. The mathematical model is: （1）； In formula (1), is the j-dimensional position value of the randomly generated i-th wave particle, that is, the random candidate solution value, A heuristic bootstrap function that adjusts the value of a random candidate solution, where To guide the strength coefficient and control the adjustment range, is a heuristic information function, which is used to represent the adjustment amount from the random candidate solution value to the center of the local area range R; the local area range R is approximated by a Gaussian distribution, and the center of the local area range R, i.e., the individual optimal position of the wave search algorithm at the tth iteration, is expressed as , the covariance matrix is ​​Σ, and the mathematical model is established to define the heuristic information function as: （2）； In formula (2), is the individual best position of the wave search algorithm at the tth iteration, is the transpose operation.

3. The method for adaptively controlling the posture of a servo robot hand platform according to claim 2, characterized in that: The heuristic guidance strategy is used in the initialization phase and local development phase of the wave search algorithm. The specific method is as follows: S411, when initializing the algorithm population position in each iteration, the guide function h(x) is used to generate a candidate solution set in the algorithm search space, and the mathematical model is: （3）； In formula (3), is the position of the i-th wave particle individual in the candidate set, is the j-dimensional position value of the randomly generated i-th wave particle individual, that is, the random candidate solution value, A heuristic bootstrap function that adjusts the values ​​of random candidate solutions; S412, in the local convergence stage, the guide function h(x) is used to update the candidate solution set, the candidate solution set is the wave particle population position, and the mathematical model is: （4）； In formula (4), is the position value of the j-th dimension of the ith wave particle in the t+1th iteration, is the position value of the j-th dimension of the ith wave particle in the t-th iteration, is the maximum number of iterations, For A heuristic guidance function to adjust the solution value; is the global optimal position of individual wave particles.

4. The method for adaptively controlling the posture of a servo robot hand platform according to claim 3, characterized in that: In S42, the waveform size control coefficient α is improved based on the dynamic sine strategy of the variance of the fitness values ​​between wave particles and the global minimum fitness value of the wave particles. The mathematical model of the improved waveform size control coefficient α(t) is: （5）； In formula (5), is the initial value of the improved waveform size control coefficient, is the variance of the fitness value, is the average fitness value, is the fitness value, is the global minimum fitness value of the wave particle, is the minimum fitness value of the tth iteration, where The mathematical model is: （6）； In formula (6), N is the size of the wave particle population, is the fitness value of the i-th wave particle position.

5. The method for adaptively controlling the posture of a servo robot arm pan / tilt platform according to claim 4, characterized in that: In said S42, the improved waveform size control coefficient improves the wave search algorithm local development phase wave particle position update mathematical model, specifically the emission electromagnetic wave mathematical model in the local development phase; （7）； In formula (7), is the new position of the ith wave particle at the tth iteration, is the current optimal wave particle position, is the current second-best wave particle position, is a column vector whose elements follow a normal distribution. is the improved waveform size control coefficient value of the tth iteration.

6. A servo robot arm pan-tilt attitude adaptive control method according to claim 5, characterized in that: The improved wave search algorithm is used to adjust the Kp, Ki, and Kd control parameters of the position PID to obtain the optimal control parameters. The specific steps are as follows: P1. Initialize the parameters of the improved wave search algorithm, including: wave particle population size N, problem dimension D, search space upper bound UB 、 The search space lower bound LB and the maximum number of iterations T max and improved waveform size control coefficients; P2, encoding the Kp, Ki, Kd proportional coefficients of the position PID algorithm into N groups of three-dimensional space vectors, and the space vectors are mapped to the real numbers of the wave particle positions; P3. Randomly initialize the initial position of the improved wave search algorithm population, and use the mathematical model of formula (3) to generate candidate solutions using the guide function h(x) in the search space of the wave search algorithm , where the upper and lower bounds of the search space are the value ranges of the three-dimensional space vector; P4. Determine whether the current number of iterations t is equal to the maximum number of iterations T max If yes, then output the wave particle position corresponding to the minimum fitness value and resolve it into the optimal Kp, Ki, Kd proportional coefficient values ​​of the position PID algorithm; otherwise, return to execute P5; P5. Establish an improved wave search algorithm position update mathematical model to update the wave particle position, including: P51, by introducing the random number r1 and the current number of iterations, a mathematical model for position update in the global search phase is established to generate the new position of the wave particle; P52, simulate the radar wave emission process, establish a mathematical model according to formula (7) to generate new wave particle positions, and explore the potential optimal solution area; P53, simulate the reflection and reception process of radar waves, establish a mathematical model for the local development stage to update the position of the wave particle population, and gradually approach the global optimal solution by updating the wave particle position through multiple iterations. The global optimal solution is the wave particle position with the smallest fitness value in the wave particle population; P54, using the mathematical model formula (4) to further update the current wave particle population position; P6. Use the objective function to calculate the fitness value of each wave particle position, retain the wave particle position corresponding to the minimum fitness value from initialization to the tth iteration, execute t=t+1 for the current iteration number t, and return to execute step P3.

7. The method for adaptively controlling the posture of a servo robot arm pan / tilt platform according to claim 1, characterized in that: The output PTZ drive value , the mathematical model is: （8）； In formula (8), For the The gimbal drive value in seconds, is the running time, For the The gimbal rotation angle error value in seconds, is the optimal ratio parameter, is the optimal integration parameter, is the optimal differentiation parameter.

Images