A method for adjusting angular velocity control of a bionic mechanical arm elbow joint

Abstract

The present application belongs to the technical field of machine arm control, and relates to a bionic mechanical arm elbow joint adjusting angular velocity control method, which solves the problem that the existing method is difficult to meet the collaborative demand of soft start, smooth transition and rapid response in bionic action, comprising: constructing a terminal two-dimensional position model with the elbow as the origin, obtaining a starting and target position to generate a motion trajectory; based on forward and inverse kinematics, deriving an expected angle sequence and calculating a target angular velocity, real-time collecting current angle and angular velocity information; constructing an angle error signal, using a position type PID controller to generate an angular velocity control instruction, and dynamically setting PID parameters through an improved state optimization algorithm; inputting the control instruction into the elbow joint driving unit to realize dynamic adjustment and closed-loop control of the angular velocity. The effect of the present application is verified through experiments, and the experiments prove that the method can significantly improve the precision, robustness and dynamic response capability of the bionic mechanical arm angular velocity control.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of robot arm control, and in particular to a bionic robot arm elbow joint adjusting angular velocity control method BACKGROUND

[0002] With the development of bionic robot technology, robot arms are increasingly widely used in medical rehabilitation, industrial cooperation, human-computer interaction and other fields. In order to achieve accurate trajectory tracking of the end effector of the robot arm, dynamic adjustment of the joint angle and angular velocity is usually required. Among them, angular velocity control as an important link connecting trajectory planning and driving execution directly affects the response speed and motion stability of the robot arm.

[0003] Traditional angular velocity control methods are mostly based on closed-loop adjustment of joint position errors, and are usually implemented by using a proportional-integral-derivative (PID) controller; in the process of simulating natural human motion, due to the nonlinear regulation characteristics and the phase response characteristics of the human muscle nervous system, the traditional PID control has the problems of large regulation rigidity, poor adaptability, uncoordinated response stages, etc., and it is difficult to meet the collaborative needs of soft start, smooth transition and fast response in bionic motion; in addition, some existing control methods lack a dynamic updating mechanism for the feedback model in the joint angular velocity control process, and cannot adjust the control instructions in real time according to different motion stages, resulting in trajectory deviation accumulation and system robustness decline; therefore, there is an urgent need for an angular velocity control method that integrates human bionic characteristics and has strong self-adaptability to improve the dynamic performance and control precision of bionic robot arms.

[0004] In order to improve the control performance, intelligent optimization algorithms are gradually introduced into the angular velocity control of robot arms; State-Based Optimization (SBO) is a new intelligent optimization method that simulates the state interaction behavior between individuals in a group to achieve fast search of global optimal solution; by introducing an elite interaction stage (global search stage) and a resource acquisition stage (local development stage), SBO algorithm jointly models the historical state of individuals, the global state of the group and environmental adaptability, thereby achieving efficient optimization in high-dimensional complex space; compared with traditional particle swarm or genetic algorithms, SBO has significant advantages in convergence speed, global search ability and avoidance of local extreme value. SUMMARY

[0005] The present application aims to solve the problems of unstable trajectory tracking, insensitive control response and poor motion stage adaptability in existing bionic robot arm joint angular velocity control methods, and proposes a bionic robot arm joint adjusting angular velocity control method that combines the characteristics of human upper limb motion, builds a two-dimensional position model at the end and a dynamic angular velocity control strategy, and realizes efficient, soft and high-precision control of joint motion.

[0006] To achieve the above-mentioned application purposes, the application adopts the following technical solutions to achieve the above-mentioned application purposes: the application relates to a bionic mechanical arm elbow joint adjusting angular velocity control method, and the specific method is as follows:

[0007] A two-dimensional position model of the end of the bionic mechanical arm elbow joint is constructed with the bionic mechanical arm elbow position as the origin, the starting position and the terminal position are obtained, and the motion trajectory data of the bionic mechanical arm is generated; the motion trajectory is composed of N1 points (x, y) in the two-dimensional position model;

[0008] According to the inverse kinematics principle, the expected angle sequence of the mechanical arm elbow joint in the trajectory is calculated, the corresponding angular velocity data is derived according to the expected angle sequence, and the current angular velocity information of the elbow joint is collected in real time;

[0009] The angular velocity error signal is constructed according to the collected current angular velocity and the target angular velocity, and the error signal is calculated through the position type PID controller to generate the angular velocity control instruction; the position type PID controller dynamically adjusts the control parameters through the improved state optimization algorithm;

[0010] The angular velocity control instruction is input into the elbow joint driving unit of the bionic mechanical arm to realize dynamic adjustment and closed-loop control of the joint angular velocity.

[0011] Among them, the bionic mechanical arm establishes a Cartesian two-dimensional position model with the elbow position as the origin; the starting position and the terminal position of the target action are obtained, the trajectory change trend in the natural movement process of the lower arm controlled by the human elbow joint is simulated, and continuous motion trajectory data are generated, the trajectory data are composed of a plurality of discrete points (x, y) in a two-dimensional plane, and the points are connected to form a curve motion trajectory; wherein the X axis extends along the horizontal direction, and the Y axis extends along the vertical direction.

[0012] Among them, knowing the starting point and target point coordinates, the initial angle is calculated according to the starting position and the origin position, and the final elbow joint angle is calculated according to the target position and the elbow joint origin position; based on the N points (x, y) in the motion trajectory, a point is selected as the target position, the angle change sequence between the target position and the starting position is calculated according to the inverse kinematics principle; the change rate of the angle difference with respect to time is calculated for the adjacent position points in the angle change sequence to obtain the angle change rate data, the angular velocity data to the target position is generated, and the angular velocity data is the target angular velocity data at this time;

[0013] The application studies the elbow joint angular velocity control problem of a bionic robot arm, and therefore only includes the expected angle sequence of the elbow joint under the trajectory, and further calculates target angular velocity data; meanwhile, the real-time angle position and angular velocity of the elbow joint are collected as feedback information, providing dynamic state input for subsequent closed-loop control.

[0014] Preferably, the bionic robot arm elbow joint is driven by a servo motor, and the movement angle of the elbow joint is calculated according to the position of the elbow joint end and the original position in the two-dimensional position model of the bionic robot arm elbow joint end, and is recorded as the movement angle data of the elbow joint at the current time; the current angular velocity information is calculated according to the difference between the current angle data and the last sampled angle data, and the current angular velocity value is calculated by using the difference method through the ratio of the two angle data before and after the sampling interval.

[0015] Preferably, the error signal is calculated by the position type PID controller to generate an angular velocity control instruction, the position type PID controller calculates the angle error signal through proportional control parameters, integral control parameters and differential control parameters, and outputs the angular velocity control instruction at the current time; the control parameters of the PID controller are dynamically adjusted by an improved state optimization algorithm; the improved state optimization algorithm constructs an evaluation function based on feedback state quantities, and realizes online optimization of the proportional control parameters, integral control parameters and differential control parameters through the adaptive adjustment strategy designed by the application, so that the control parameters of the calculated angular velocity control instruction reach the best precision, thereby enhancing the robustness and adjustment response ability of the bionic robot arm elbow joint control system to dynamic disturbances.

[0016] Preferably, the adaptive adjustment strategy includes an elite emergence jump mechanism based on position fluctuation entropy and elite degradation index and a dimension perception disturbance updating method.

[0017] Preferably, the control parameters of the PID controller are dynamically adjusted by an improved state optimization algorithm, and the specific method is as follows:

[0018] The position vector of the individual is mapped with the control parameter vector of the PID controller, wherein the control parameters of the PID controller include proportional control parameters Kp, integral control parameters Ki and differential control parameters Kd, and each set of control parameters constructs a three-dimensional vector, after mapping, the first dimension value of the three-dimensional vector is the first dimension value of the individual position vector ; the fitness value of each agent individual position vector is calculated by using the evaluation function, the fitness values corresponding to each agent individual are sequentially sorted, and the agent individual position vector value corresponding to the minimum fitness value is taken as the global optimal solution, and the first optimal individual in the sorting is taken as the elite set.

[0019] Random initialization improves the optimization algorithm optimization process of individual position, the first dimension value Random assignment, set the value range between the maximum value ub and the minimum value lb;

[0020] Design an elite emergence jumping mechanism based on status fluctuation entropy and elite degradation index to realize the update and jump of individuals in the new elite direction and update the individual position; and construct a position update strategy in the resource acquisition stage by a dimension perception disturbance update method to realize the position update of individuals in the resource acquisition stage;

[0021] The response results of the dynamic adjustment effect of the updated proportional control parameter Kp, integral control parameter Ki and differential control parameter Kd value on the joint angular velocity are calculated by the evaluation function value, which is recorded as the updated fitness value, and then the individual position is updated by the elite emergence jumping mechanism and the dimension perception disturbance update method. The iteration number is increased once each time, until the current iteration number reaches the maximum iteration number, then the dynamic optimization is ended, and the value of the proportional control parameter Kp, the integral control parameter Ki and the differential control parameter Kd of the best PID controller is output, which is recorded as the best PID controller parameter combination.

[0022] Preferably, the evaluation function is designed by integral absolute error, and the integral absolute error (IAE) in the target response process is calculated according to the deviation value between the actual elbow joint adjustment angular velocity and the target elbow joint adjustment angular velocity of the controlled object in the control period, and the IAE is taken as the individual fitness value.

[0023] Preferably, the best PID controller parameter combination is reconstructed to calculate the error signal of the position type PID controller, generate the angular velocity control command, input the elbow joint driving unit of the bionic robot arm, and realize the dynamic adjustment and closed-loop control of the joint angular velocity by adjusting the servo actuator through the elbow joint driving unit. Specifically, the angular velocity control process of the elbow joint of the bionic robot arm is based on the dynamic response model of the elbow joint driving unit of the bionic robot arm, which takes the angular velocity control command generated by the position type PID controller as the input driving voltage signal, and acts on the servo actuator to dynamically control the actuator current. The actuator current is converted into motor output torque through the motor torque constant, and the output torque drives the elbow joint of the robot arm to produce angular velocity change around the rotation axis; finally, a response mathematical model between the control signal and the angular velocity is constructed based on the second-order dynamic response model, and the elbow joint adjustment angular velocity of the bionic robot arm is output.

[0024] Preferably, the elite emergence jumping mechanism based on status fluctuation entropy and elite degradation index is specifically as follows:

[0025] Extract the individual In the near local state change data sequence in the next iteration cycle , the distance between adjacent states is calculated, the state fluctuation entropy index of the individual is constructed, and the weighted coefficient is constructed based on the relative fluctuation intensity of each iteration , the distance and the weighted coefficient are input to construct the state fluctuation entropy index of the individual ;

[0026] An elite individual is randomly selected from the elite set, and the state fluctuation entropy of the associated elite individual is calculated according to the state fluctuation entropy index of the individual , and the state difference factor is calculated based on the state fluctuation entropy index of the individual and the associated elite individual ;

[0027] The trajectory vector is constructed based on the historical positions of the current individual in the current and the last two iteration cycles , the current evolution curvature factor of the individual is calculated according to the direction angle and the path length of the trajectory vector ; the attraction weight factor of the individual to the elite individual is established by using a nonlinear function form according to the current evolution curvature factor of the individual and combining the curvature response adjustment factor , which is used to control the degree of guidance received by the individual at the current position;

[0028] The state of each elite individual is monitored, and the number of iterations without obtaining positive evolution improvement is counted ; the ratio of the iteration number to the preset maximum tolerance threshold is calculated to obtain a trend index for judging whether to degenerate , if the trend index of any elite individual meets the degradation condition, the elite individual with the most different state from the current elite individual is selected as the jump target and its new jump elite index is constructed according to the state difference factor of the associated individual on the premise of excluding itself;

[0029] According to the elite index and the attraction weight factor, the position vector of the current individual is updated by combining the standard disturbance structure, so as to realize the updated jump of the individual in the new elite direction.

[0030] ​​​Preferably, the original resource acquisition stage position update strategy constructs a mask variable vector based on a preset number of dimensions, and performs position update based on mask values; the generation mode of the mask variable completely depends on a random function, and the influence degree of each variable dimension, i.e., the proportional control parameter Kp, the integral control parameter Ki and the differential control parameter Kd, on the optimization result of the evaluation function is not considered, resulting in that the resource fusion update lacks pertinence; based on this, a dimension-aware disturbance update method is proposed, and target-oriented disturbance is performed on each dimension value to strengthen the difference between the dimension values.

[0031] Preferably, the dimension-aware disturbance update method specifically includes:

[0032] According to the current individual In the sensitivity of the evaluation function value change in the previous optimization process, the dimension contribution index of each variable dimension is constructed The dimension contribution index is the contribution degree of each variable dimension value to the dynamic adjustment effect value of the joint angular velocity, wherein the effect value is calculated by the evaluation function feedback;

[0033] According to the dimension contribution index, a modulation factor is constructed , and the value of each dimension of the current individual is updated as follows to realize the dimension-aware disturbance update, the modulation factor is proportional to the dimension contribution index and is used to adjust the adjustment direction of the dimension value; wherein the mathematical model of the dimension-aware disturbance update is:

[0034] ;

[0035] wherein, is the value of the jth dimension of the updated ith individual position, is the value of the jth dimension of the current ith individual position, is the global optimal solution of the jth dimension in the current population.

[0036] Preferably, in the process of controlling the elbow joint angular velocity of the robot arm, the response sensitivity of each dimension of the control parameter in different control processes is obviously different, which is manifested as that some dimension disturbances have a significant promoting effect on error convergence, while some other dimensions show unstable behaviors such as disturbance reversal and interference amplification, therefore, the dimension contribution index is introduced to quantify the dynamic contribution degree of each dimension in the historical adjustment process, and further analysis is performed, by calculating the partial derivative value of different dimensions in the historical function change, the dimension contribution index can identify the main control parameters sensitive to the error in the current stage, so as to give a greater adjustment weight in the disturbance update.

[0037] Preferably, the modulation factor is calculated for each individual dimension, the amount of position change of the dimension in M consecutive iterations , and the unit direction of the dimension is defined according to the size of the change in M consecutive steps , the direction consistency index is constructed based on the consistency integral of the disturbance direction by evaluating the unit direction change of the dimension in the historical iteration , wherein the direction consistency index of each individual is valued between [-1, 1], and the modulation factor is generated according to the dimension contribution index of the target individual in the dimension and the calculated direction consistency index ; wherein, is the modulation factor of the ith individual in the tth iteration, is the dimension contribution index of the ith individual in the tth iteration.

[0038] Preferably, the evaluation function is designed with integral absolute error, and the integral absolute error (IAE) in the target response process is calculated according to the deviation value between the actual elbow joint adjustment angular velocity and the target elbow joint adjustment angular velocity of the controlled object in the control period, and the IAE is taken as the individual fitness value.

[0039] Based on the above invention content, the beneficial effects of the present application are summarized as follows:

[0040] Firstly, by constructing a two-dimensional position model of the elbow joint, combining forward and inverse kinematics to derive the expected angle and angular velocity sequence, simulating the natural motion characteristics of the human elbow, and generating a continuous and smooth control target trajectory, the flexibility and fidelity of the bionic motion are effectively improved, which is suitable for rehabilitation training, human-machine cooperation and other scenes.

[0041] Secondly, an elite emergence jump mechanism based on position fluctuation entropy and elite degeneration index is proposed, and combined with the dimension perception disturbance update method, the adaptive update of the PID controller parameters is realized; by adaptively adjusting the controller parameters, the problem of insensitive angular velocity response in the bionic robot elbow joint motion process caused by fixed parameters is effectively avoided, so that the angular velocity control process has strong dynamic response ability and anti-interference performance.

[0042] Finally, the angular velocity control command is used to drive the current change of the servo actuator, and the torque is converted through the motor torque constant to drive the elbow joint angular velocity change, and a second-order dynamic response mathematical model from the PID control signal to the angular velocity output is constructed, which ensures that the control command can be accurately and real-time mapped to the bionic robot elbow joint angular velocity control. BRIEF DESCRIPTION OF DRAWINGS

[0043] Figure 1 is the implementation flowchart of the bionic robot elbow joint adjustment angular velocity control method of the present application.

[0044] Figure 2 A structural block diagram of a bionic mechanical arm elbow joint adjusting angular velocity control method of the application.

[0045] Figure 3 A simulation structural diagram of the bionic mechanical arm elbow joint adjusting angular velocity control of the application.

[0046] Figure 4 A fitness diagram of the bionic mechanical arm elbow joint adjusting angular velocity control of the application.

[0047] Figure 5 A control parameter optimization result diagram of the elbow joint position type PID control algorithm of the improved front control algorithm.

[0048] Figure 6 A control parameter optimization result diagram of the elbow joint position type PID control algorithm of the improved rear control algorithm.

[0049] Figure 7 A control effect diagram of a bionic mechanical arm elbow joint adjusting angular velocity control method of the application. DETAILED DESCRIPTION

[0050] The technical solutions in the embodiments of the application will be clearly and completely described below with reference to the drawings in the embodiments of the application. Obviously, the described embodiments are only some of the embodiments of the application but not all the embodiments of the application.

[0051] Based on the embodiments in the application, all other embodiments obtained by those skilled in the art without creative labor are within the protection scope of the application. In the description of the application, it should be understood that the terms "center", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer" and the like indicate the orientation or positional relationship shown in the drawings, and are only used to facilitate the description of the application and simplify the description, and therefore cannot be understood as indicating or implying that the device or element referred to must have a particular orientation, be constructed and operated in a particular orientation, and therefore cannot be understood as a limitation on the application.

[0052] In order to realize accurate control of the angular velocity of the elbow joint of the bionic mechanical arm, the application relates to a bionic mechanical arm elbow joint adjusting angular velocity control method, which is described in detail as follows. Figure 1

[0053] S1, a two-dimensional position model of the end of the elbow joint of the bionic mechanical arm is constructed with the elbow position of the bionic mechanical arm as the origin, the starting position and the terminal position are obtained, and the motion trajectory data of the bionic mechanical arm is generated; the motion trajectory is composed of N points (x, y) in the two-dimensional position model;​

[0054] S2, according to the principle of inverse kinematics, calculating the expected angle sequence of the elbow joint of the robotic arm in the trajectory, deriving the corresponding angular velocity data according to the expected angle sequence, and collecting the current angular velocity of the elbow joint in real time;

[0055] S3, constructing an angular velocity error signal according to the collected current angular velocity and the target angular velocity, and calculating the error signal through a positional PID controller to generate an angular velocity control instruction; the positional PID controller dynamically adjusts the control parameters through an improved state optimization algorithm;

[0056] S4, inputting the angular velocity control instruction to the elbow joint driving unit of the bionic robotic arm to realize dynamic adjustment and closed-loop control of the joint angular velocity.

[0057] In some embodiments of the present application, the bionic robotic arm elbow joint angular velocity control system includes a trajectory planning module, a joint angle solving module, a sensing acquisition module, an angular velocity controller module, and an elbow joint driving unit.

[0058] In embodiments of the present application, the trajectory planning module is configured to construct a two-dimensional position model of the elbow joint end effector according to the set starting position and target position, specifically, the bionic robotic arm establishes a Cartesian two-dimensional position model with the elbow position as the origin (0, 0); the starting position of the action is obtained and the end position , where N1=1, 2, 3,..., 100, by simulating the trajectory change trend in the natural movement process of the lower arm controlled by the human elbow joint, continuous motion trajectory data is generated, the trajectory data is composed of 100 discrete points in a two-dimensional plane, and these points form a curve motion trajectory; wherein the X-axis extends along the horizontal direction and the Y-axis extends along the vertical direction.

[0059] In embodiments of the present application, based on the N points (x, y) in the motion trajectory, a point is selected as the target position, and the angle change sequence between the target position and the starting position is calculated; the change rate of the angle difference with respect to time is calculated for adjacent position points in the angle change sequence to obtain the angle change rate data, and the angular velocity data reaching the target position is generated, at which time the angular velocity data is the target angular velocity data;

[0060] In the present application, the expected angle sequence in the trajectory is calculated, and the expected elbow joint angle sequence is derived based on inverse kinematics in the implementation process, and the mathematical model is:

[0061] ;

[0062] wherein, and are the horizontal and vertical position coordinates of the end effector of the bionic robot arm in a two-dimensional plane, k = 1, 2, 3,..., N, which are randomly selected; and are the starting position coordinates of the end effector of the bionic robot arm in the horizontal and vertical directions in a two-dimensional plane;

[0063] According to the expected angle sequence, a corresponding angular velocity data sequence is derived, and the mathematical model is:

[0064] ;

[0065] wherein, is the angular velocity corresponding to the target angle when the kth position is the target position coordinate; is the angle between the starting position and the elbow joint origin position, and the kth position in the expected elbow joint angle sequence is the target position coordinate; is the time from the starting position to the target position, is the target angle corresponding to the target position coordinate of the kth discrete point coordinate in the expected elbow joint angle sequence;

[0066] In the embodiment of the present application, the target position is a discrete point selected from continuous motion trajectory data (curved motion trajectory), denoted as , the time from the initial position to the target position is set as , and the data is brought into the angular velocity data sequence to obtain the target angular velocity value of the present embodiment .

[0067] The current angular velocity data of the elbow joint is collected in real time by the sensing acquisition module to the error calculation unit of the angular velocity controller module, and the error with the target angular velocity value is recalculated based on the current data ; the angular velocity controller module includes an error calculation unit, a position type PID control algorithm unit and an angular velocity control instruction output unit, the error between the target angle and the real-time angle is calculated, and the angular velocity control instruction is calculated through the improved position type PID control algorithm, and the control parameters in the position type PID controller are dynamically self-adaptively adjusted through the improved state optimization algorithm.

[0068] As described above, the PID parameters include a proportional parameter Kp, an integral parameter Ki iAnd the differential parameter Kd, the improved state optimization algorithm of the application is an improvement based on the standard state optimization algorithm, which mainly involves the improvement of the position updating strategy, the control parameters of the PID controller described in the application are dynamically adjusted by the improved state optimization algorithm; The improved state optimization algorithm constructs an evaluation function based on the feedback state quantity, and realizes online optimization of the proportional control parameter, the integral control parameter and the differential control parameter through the adaptive adjustment strategy designed in the application, so that the control parameters of the calculation angular velocity control instruction reach the best precision.

[0069] Please refer to Figure 2 The process of the improved state optimization algorithm for online optimization and adjustment of the PID parameters of the speed PID controller is specifically described.

[0070] S31, the position vector of the individual is mapped with the control parameter vector of the PID controller; wherein the control parameters of the PID controller include proportional control parameters Kp, integral control parameters Ki and differential control parameters Kd, each set of control parameters constructs a three-dimensional vector, and after mapping, the first dimension value of the three-dimensional vector is the first dimension value of the individual position vector , wherein .

[0071] S32, set the initialization parameters of the improved state optimization algorithm; the initialization parameters include: the maximum iteration number T, the problem dimension D, the upper limit position vector ub and the lower limit position vector lb of the solution space optimized by the algorithm, and the population size N;

[0072] In this embodiment, the maximum iteration number T is set to 80, the problem dimension D is set to 3, the upper limit position vector ub of the solution space optimized by the algorithm is set to [90 90 90], the lower limit position vector lb is set to [0.01 0.01 0.01], and the population size N is set to 20;

[0073] Step S32 is used for the initialization stage of the improved state optimization algorithm, including generating each individual position between the upper limit position vector ub and the lower limit position vector lb of the solution space by using a random method, so as to provide an initial value for subsequent individual position updating;

[0074] In the initialization stage, the population individual position is randomly generated in the search space:

[0075] , wherein is the jth dimension value of the ith individual, and the whole is the individual position vector, and is the upper limit position vector and the lower limit position vector of the ith individual in the solution space, and rand is a random number between 0 and 1.

[0076] S33, calculate the fitness value of each agent individual position vector using the evaluation function, sequentially sort the fitness values corresponding to each agent individual, and take the agent individual position vector value corresponding to the minimum fitness value as the global optimal solution, and the top optimal individuals in the sorting as the elite set; in this embodiment, the top 5 optimal individuals are selected as the elite set;

[0077] The evaluation function is used to evaluate the accuracy of the solution corresponding to each individual position and the change speed of the evaluation function value to evaluate the optimization speed of the solution. The smaller the fitness value calculated by the evaluation function, the higher the accuracy of the solution, and the faster the change speed of the fitness value, the faster the optimization speed of the solution.

[0078] As described above, the application gives an evaluation function model. The evaluation function adopts the design of integral absolute error, calculates the integral absolute error (IAE) in the target response process according to the deviation value between the actual elbow joint adjustment angular velocity and the target elbow joint adjustment angular velocity of the controlled object in the control period, and the mathematical model is: wherein is the total running time, and the simulation time in siumulink is 60 seconds.

[0079] S34, reconstruct the global search and local development stages of the state optimization algorithm based on the elite emergence jump mechanism and the dimension perception disturbance update method of the status fluctuation entropy and the elite degradation index to obtain an improved state optimization algorithm, and update the position vector of the individual;

[0080] The position update strategy of the elite contact stage and the resource acquisition stage is constructed to realize the construction of the global search strategy and the local development strategy of the control parameter optimization of the PID controller. Based on the above, the elite emergence jump mechanism based on the status fluctuation entropy and the elite degradation index proposed in the elite contact stage is taken as the global search strategy, and the dimension perception disturbance update method proposed in the resource acquisition stage is taken as the local development strategy.

[0081] The global search strategy includes constructing the status fluctuation entropy index , calculating the state difference factor , establishing the attraction weight factor , and designing the trend index to determine whether it is degraded to realize the update jump of the individual in the new elite direction.

[0082] In this embodiment, the local state change data sequence of the individual in the last iteration period is extracted, the distance between adjacent states is calculated, and the weighted coefficient is constructed based on the relative fluctuation intensity of each iteration, ;

[0083] in, and Let i be the position vector of the i-th individual in the t-th and t-1-th iterations. To find the minimum value, avoid the denominator being zero;

[0084] Using the distance and weighting coefficients as input, an individual is constructed. State fluctuation entropy index

[0085] ;

[0086] In some embodiments of this application, Set to 0.00001.

[0087] In this embodiment, an elite individual is randomly selected from the elite set, and based on the individual... The state fluctuation entropy index is used to calculate the elite individuals associated with the individual. State fluctuation entropy Based on the state fluctuation entropy index of the individual and the associated elite individuals, a state difference factor is calculated. .

[0088] In this embodiment, the current individual is obtained. Construct trajectory vectors from historical positions in the current and two previous iteration cycles. The current evolution curvature factor of the individual is calculated based on the directional angle of the trajectory vector and the path length. The individual's current evolutionary curvature factor is designed based on the discrete curvature principle; the mathematical model is as follows:

[0089] ;in, Not equal to ;

[0090] Based on the individual's current evolutionary curvature factor and combined with the curvature response adjustment factor, an attraction weighting factor for the individual to elite individuals is established using a nonlinear function. The attraction weighting factor is used to control the degree of guidance an individual receives at their current location, and whether it enhances their dependence on historical elites. The mathematical model is as follows:

[0091] ;

[0092] in, This is the curvature response adjustment factor. In this embodiment, its value must be greater than zero. The specific value needs to be adjusted based on the effect of the attraction weighting factor. The initial setting is 0.5. For the set curvature reference, the individual current evolution curvature factor is manually adjusted, and for the stability boundary threshold, the value is within 0 to 1, and in the embodiment, it is set to 0.2, The evolution curvature factor of the i-th individual in the trajectory of the last three generations;

[0093] When the individual evolution curvature factor is larger, it indicates that the trajectory direction of the position update is unstable, and the individual position update tends to depend on the elite individual , the attraction weight factor tends to 1, and on the contrary, when the individual evolution curvature factor is smaller, it tends to 0, which indicates that the evolution direction of the position update is stable, and the individual position update tends to depend on the individual position, and the attraction weight factor tends to 0.

[0094] In the embodiment, the state of each elite individual is monitored, and the number of iterations of continuous non-positive evolution improvement is counted ; the number of iterations is calculated by the ratio of the preset maximum tolerance threshold , and a trend index for judging whether to degrade is obtained ; wherein the number of iterations of continuous non-improvement of the elite individual position is judged by comparing the fitness value of the individual in each iteration, is a fixed set parameter, indicating the maximum iteration step number allowed for the elite individual to maintain the current state, and is set to 10;

[0095] If the trend index of any elite individual satisfies the degradation greater than condition;

[0096] The embodiment introduces an elite jump mechanism to break local convergence. According to the state difference factor of the associated individual, the elite individual with the most different state from the current elite individual is selected as the jump target under the premise of excluding itself, and a new jump elite index is constructed accordingly , and the mathematical model is: .

[0097] In the embodiment, according to the elite index and the attraction weight factor, the position vector of the current individual is updated in combination with the standard disturbance structure, so as to realize the update jump of the individual in the new elite direction;

[0098] ;

[0099] Wherein, is the updated i-th individual position vector, is a standard normal distribution with a mean of 0 and a variance of 1, which is used to disturb the individual position and reduce the probability of the individual position update falling into local optimum.

[0100] The local development strategy includes a dimension-aware perturbation update method, which calculates a dimension contribution index , constructs a modulation factor , realizes dimension-aware oriented perturbation update, and updates a certain dimension of the individual position vector.

[0101] In this embodiment, according to the current individual , the sensitivity of the evaluation function value change in the optimization process is constructed for each variable dimension , the dimension contribution index , the contribution of each variable dimension value to the dynamic adjustment effect value of the joint angle velocity, wherein the effect value is calculated by the evaluation function feedback; wherein the mathematical model of the dimension contribution index is:

[0102] ;

[0103] wherein, is the maximum number of iterations, is the individual fitness value of the tth iteration, is the jth dimension value of the individual of the tth iteration.

[0104] In this embodiment, according to the dimension contribution index, a modulation factor is constructed, and the value of each dimension of the current individual is updated as follows to realize dimension-aware oriented perturbation update, the modulation factor is proportional to the dimension contribution index, and is used to adjust the adjustment direction of the dimension value; wherein the mathematical model of the dimension-aware oriented perturbation update is:

[0105] ;

[0106] wherein, is the jth dimension value of the updated ith individual position, is the jth dimension value of the current ith individual position, is the global optimal solution of the jth dimension in the current population.

[0107] Wherein, the modulation factor is calculated for each dimension of each individual, the position change amount of the dimension in M consecutive iterations, and the unit direction of the dimension is defined according to the size of the change amount in M consecutive steps, the unit direction change of the dimension in the history iteration is evaluated, and the direction consistency index is constructed based on the consistency integral of the perturbation direction; wherein the mathematical model of the direction consistency index of each individual is:

[0108] ; wherein is the direction consistency index of the ith individual in the tth iteration, is the unit direction of the dimension of the ith individual in the tth iteration, is the unit direction of the dimension of the ith individual in the t-1th iteration; the direction consistency index of each individual is valued between [-1, 1], and the adjustment factor is generated according to the dimension contribution index of the target individual in the dimension and the calculated direction consistency index ; wherein is the modulation factor of the ith individual in the tth iteration, is the dimension contribution index of the ith individual in the tth iteration.

[0109] S35, after the position is updated, the fitness value is recalculated, and the best individual position vector value is located , after the number of updates reaches the maximum iteration number, the individual position vector corresponding to the minimum fitness value is output, and the global best proportional control parameter Kp, integral control parameter Ki and differential control parameter Kd are obtained.

[0110] In this embodiment, the values of the global best proportional control parameter Kp, integral control parameter Ki and differential control parameter Kd are used to calculate the error signal to generate the angular velocity control instruction , specifically, the first dimension value of the individual position vector corresponding to the minimum fitness value is assigned to the proportional control parameter Kp, the second dimension value is assigned to the proportional control parameter Ki, and the third dimension value is assigned to the proportional control parameter Kd, and the mathematical model of the angular velocity control instruction generated by the position PID control algorithm is:

[0111] ;

[0112] wherein, is the angular velocity control instruction of the current control period; is the angular velocity error value of the current control period, is the angular velocity error value of the last control period.

[0113] The elbow joint driving unit is used to receive the angular velocity control instruction, adjust the working state of the servo actuator, drive the elbow joint to perform corresponding motion at the expected angular velocity, realize closed-loop tracking control, and the physical model of the elbow joint is the servo actuator. The angular velocity control process of the bionic robot elbow joint is based on the servo motor driving mechanism to construct a dynamic response model;

[0114] In this embodiment, the bionic mechanical arm elbow joint uses a small servo motor control, through the angular velocity control instruction control driver to realize the torque control of servo motor, so as to realize the control of angular velocity, this embodiment establishes the servo drive model, that is, the driving model between the angular velocity of elbow joint (servo motor) and angular velocity control instruction.

[0115] The dynamic response model establishes a second-order dynamic response equation based on Kirchhoff's law, and the mathematical model is:

[0116] ;

[0117] Among them, is the angular velocity control instruction of the current control period, is the armature inductance, is the armature resistance, is the back electromotive force constant, is the current angular velocity rad / s of the current control period.

[0118] In this embodiment, the armature inductance L = 0.005H, the armature resistance R = 1.0Ω, the back electromotive force constant is 0.05V·s / rad;

[0119] According to the moment of inertia and the damping characteristics, the torque- angular velocity dynamic relationship is constructed, and the dynamic relationship between torque and angular velocity is established, and the mathematical model is:

[0120] ;

[0121] Among them, is the moment of inertia, is the viscous damping coefficient, is the current output torque; in this embodiment, the moment of inertia J = 0.01kg·m², the viscous damping coefficient B = 0.02N·m·s / rad; the linear relationship between current and torque is substituted to establish the second-order dynamic response equation of the control signal to the angular velocity;

[0122] ;

[0123] Among them, is the torque constant, in this embodiment, the torque constant = 0.05N·m / A;

[0124] The second-order dynamic response equation is obtained.

[0125] In this embodiment, the controlled function is controlled by using the improved control algorithm, the improved control algorithm and the position type PID control algorithm, the angular velocity value is set to 9 rad / s, that is, the angular velocity has reached the steady state, and the derivative term is zero, then according to the above, the angular velocity control instruction of the current control period should be 4V, the angular velocity value is set to 31 rad / s, and the angular velocity control instruction of the current control period should be 14V, the total running time is 60 seconds, please refer to Figure 3 The second-order dynamic response equation in the complex frequency domain is shown in the following formula As the input of the elbow joint angular velocity control system of the bionic robot arm, the elbow joint driving unit of the Siumulink simulation model of the controlled function is run, and the elbow joint angular velocity control system of the bionic robot arm is obtained as shown in the following figure Figure 4 to Figure 7 The method of the application has the above effects.

[0126] Further, please refer to Figure 3 and Figure 4 As shown in the following figure Figure 3 The fitness value monitoring part in the above formula, the fitness value of the improved control algorithm of the application for the control parameter optimization of the elbow joint position type PID control algorithm is 12.314; the fitness value of the improved control algorithm before and after optimization is Figure 4 The fitness value of the improved control algorithm before and after optimization is 81.7029; the smaller the fitness value, the higher the control accuracy, and the higher the control parameter optimization of the elbow joint position type PID control algorithm by the improved control algorithm.

[0127] Further, please refer to Figure 5 and Figure 6 As shown in the following figure In this embodiment, the Kp of the control parameter optimization of the elbow joint position type PID control algorithm by the improved control algorithm is 13.1784, the Ki is 11.3693, and the Kd is 0.01; the Kp of the control parameter optimization of the elbow joint position type PID control algorithm by the improved control algorithm is 0.856797, the Ki is 1.15773, and the Kd is 0.274371.

[0128] Figure 7As shown, the Siumulink simulation model sets the total running time to 60 seconds, and the angular velocity control instruction is monitored. It can be obviously seen from the control effect that the improved control algorithm of the application is superior to the control algorithm before improvement and the position type PID control algorithm in the accuracy and control speed of the elbow joint angular velocity control of the bionic robot arm. The improved control algorithm of the application does not have overshoot in the elbow joint angular velocity control of the bionic robot arm, but the control algorithm before improvement and the position type PID control algorithm have large overshoots, respectively, and the improved control algorithm has a shorter steady-state time in the elbow joint angular velocity control of the bionic robot arm. It is proved that the method can significantly improve the accuracy, robustness and dynamic response ability of the bionic robot arm angular velocity control.

Claims

1. A method for adjusting the angular velocity of a bionic robotic arm elbow joint, characterized in that, A two-dimensional position model of the end of the elbow joint of the bionic robot arm is constructed with the elbow position of the bionic robot arm as the origin, the starting position and the end position are obtained, and the motion trajectory data of the bionic robot arm is generated; the motion trajectory is composed of N1 points (x, y) in the two-dimensional position model; According to the principle of inverse kinematics, the expected angle sequence of the elbow joint of the robot arm in the trajectory is calculated, the corresponding angular velocity data is derived according to the expected angle sequence, and the current angular velocity information of the elbow joint is collected in real time; An angular velocity error signal is constructed according to the collected current angular velocity and target angular velocity, and an error signal is calculated by a position PID controller to generate an angular velocity control instruction; the position PID controller dynamically adjusts the control parameters through an improved state optimization algorithm; The position PID includes proportional parameter Kp, integral parameter Ki and differential parameter Kd, the improved state optimization algorithm constructs an evaluation function based on feedback state quantity, and realizes online optimization of proportional control parameter, integral control parameter and differential control parameter through adaptive adjustment strategy; the feedback state quantity is the current angular velocity; The evaluation function adopts a whole integral absolute error design, calculates the integral absolute error in the target response process according to the deviation value between the actual elbow adjustment angular velocity and the target elbow adjustment angular velocity in the control period. The adaptive adjustment strategy improves the position update strategy of the global search and local development stage of the state optimization algorithm, and the adaptive adjustment strategy includes an elite emergence jumping mechanism based on position fluctuation entropy and elite degradation index and a dimension perception disturbance updating method; The angular velocity control instruction is input to the elbow joint driving unit of the bionic robot arm to realize dynamic adjustment and closed-loop control of the joint angular velocity.

2. The method according to claim 1, wherein The two-dimensional position model is a Cartesian two-dimensional position model, which simulates the trajectory change trend in the natural movement process of the lower arm controlled by the human elbow joint, takes the starting position and the end position of the target action as two endpoints, and moves with the arm length as the radius to form a sector motion trajectory data, the arm length is the length from the elbow position of the bionic robot arm to the end of the elbow joint of the robot arm.

3. The method of claim 2, wherein the method further comprises: determining a target angular velocity of the elbow joint; and adjusting the target angular velocity of the elbow joint based on the target angular velocity and the current angular velocity of the elbow joint. The desired angle sequence is the angle of each point (x, y) on the motion trajectory in the two-dimensional position model relative to the starting position of the clip angle; The corresponding angular velocity data is obtained by: selecting a point as a target position based on N points (x, y) in the motion trajectory, calculating the angle change sequence between the target position and the starting position according to the principle of inverse kinematics; the change rate of the angle difference with respect to time is calculated for adjacent position points in the angle change sequence to obtain the angle change rate data, and the angular velocity data reaching the target position is generated, at this time the angular velocity data is the target angular velocity data.

4. The method of claim 1, wherein the method further comprises: The elite emergence jumping mechanism based on position fluctuation entropy and elite degradation index improves the global search stage of the state optimization algorithm, and the specific method is: extracting individuals in the vicinity local state change data sequence in the next iteration cycle , calculate the distance between adjacent states, construct the state fluctuation entropy index of the individual, and construct the weighted coefficient based on the relative fluctuation intensity of each iteration , taking the distance and the weighted coefficient as input, constructing the state fluctuation entropy index of the individual ;​ Randomly select an elite individual from the elite set, and then... The state fluctuation entropy index is used to calculate the elite individuals associated with the individual. State fluctuation entropy And calculate the state difference factor based on the state fluctuation entropy index of the individual and the associated elite individuals. ; acquiring a current individual constructing a trajectory vector based on the historical positions in the current and previous two iteration periods calculating a current evolution curvature factor of the individual according to the included angle of the trajectory vector and the path length establishing an attraction weight factor of the individual to the elite individual by using a nonlinear function form according to the current evolution curvature factor of the individual and in combination with a curvature response adjustment factor the attraction weight factor is used to control the degree of guidance on the individual at the current position for each elite individual state monitoring, counting the number of iterations of consecutive non-positive evolutionary improvement ; compare the number of iterations with a preset maximum tolerance threshold ratio calculation, obtaining a trend index for judging whether to degrade , if the trend index of any elite individual meets the degradation condition, according to the state difference factor of the associated individual, excluding itself, selecting the elite individual with the most different state from the current elite individual as the jump target, and constructing a new jump elite index accordingly; Based on the elite index and attraction weight factor, combined with the standard normal distribution, through elite individuals For the current individual The relative position of the position vector is updated to realize the position update of the individual under the new elite direction.

5. The method of claim 4, wherein the method further comprises: The dimension perception disturbance updating method improves the local development stage of the state optimization algorithm, including: Based on the current individual The sensitivity of the evaluation function value to changes during each optimization process is assessed, and each variable dimension is constructed. Dimensional contribution metrics The dimensional contribution index is the contribution of each variable dimension value to the dynamic adjustment effect value of the joint angular velocity, wherein the effect value is calculated and fed back through an evaluation function; According to the dimension contribution index, a modulation factor is constructed And the value of the current individual in each dimension The following update is performed, a dimension-aware guiding disturbance update, the modulation factor is proportional to the dimension contribution index, and is used to adjust the adjustment direction of the dimension value; wherein the mathematical model of the dimension-aware guiding disturbance update is: ； wherein, is the value of the jth dimension of the ith individual position after update, is the value of the jth dimension of the ith individual position at present, is the global optimal solution of the jth dimension in the current population.

6. The method of claim 5, wherein the method further comprises: The dynamic adjustment of the control parameters includes: mapping the position vector of the individual to a control parameter vector of the PID controller; wherein the control parameters of the PID controller comprise a proportional control parameter Kp, an integral control parameter Ki and a differential control parameter Kd, each set of control parameters forming a three-dimensional vector, and wherein after the mapping the first dimension value of the three-dimensional vector is the first value of the position vector of the individual ;​​​ The initialization parameters of the improved state optimization algorithm are set; the initialization parameters include: the maximum number of iterations T, the problem dimension D, the upper limit position vector ub and the lower limit position vector lb of the solution space of the algorithm optimization, and the population size N; The fitness value of each agent individual position vector is calculated by using the evaluation function, the corresponding fitness value of each agent individual is sequentially sorted, the agent individual position vector value corresponding to the minimum fitness value is taken as the global optimal solution, and the top optimal individuals in the sorting are taken as the elite set; The improved state optimization algorithm is obtained by reconstructing the global search and local development stages of the state optimization algorithm based on an elite emergence jump mechanism and a dimension perception disturbance updating method based on status fluctuation entropy and elite degeneration index, and the position vector of the individual is updated; After the number of updates reaches the maximum iteration number, the position vector of the individual corresponding to the minimum fitness value is output, and the global optimal proportional control parameter Kp, integral control parameter Ki and differential control parameter Kd are obtained by analysis.

Images