Computational Intelligence GWO Algorithm Optimization And Its Application In Motion Control

The industrial production and science technology are increasingly involved with motion control.Permanent Magnet Synchronous Motor(PMSM),with its superior performance in motion control system,has played an important role in all walks of life.Robot,as a powerful substitute for human labor,is serving many industries.Whether it is a single PMSM or a complex robot,the motion control is a complex system with nonlinear,strong coupling and multiple variables.The traditional optimization can hardly satisfy the requirements of modern motion control,while the computational intelligence has the great advantages of self-adaptability and robustness,providing a new means for the optimal control of the complex nonlinear motion system.This study mainly investigated the optimization of computational intelligence algorithm and its application in motion control system,involving algorithm improvement and its application.The first step is the study of GWO and its improvement.Firstly,the introduction and analysis of the GWO have been made,followed by the verification of its global convergence.To increase the diversity of samples and reduce the probability of GWO falling into the local optimal,an optimized GWO based on eliminating-reconstructing mechanism and mutation operator has been proposed.Furthermore,by giving the derivation algorithm,it comes to the conclusion that the eliminating-reconstructing mechanism and the mutation operator complement each other in the function.The standard GWO was seen as the static GWO.To reduce the waiting time of the search wolves,two dynamic GWO structures have been proposed.In the dynamic GWOs,there is no need to wait for the other wolves,which has improved the iterative convergence speed and kept the algorithms more competitive.Based on the structure of the dynamic GWOs,the performance of the other improved GWOs have been probed.It has further verified that the performance of the improved dynamic-GWO-structure algorithms were better than the static-GWO-structure algorithms.The second step focuses on the application of GWO and other computational intelligence algorithms to the control and synchronous control of the PMAM chaos.When PMSM was unfavorable,a chaos controller has been proposed on the basis of Hamiltonian theory and GWO,which was composed of the nonlinear disturbance compensation and the tracking controller.After analyzing the nonlinear disturbance in this strictly dissipative generalize reduction model,a nonlinear disturbance compensator with adjustable gain was proposed.Under the Lyapunov theorem,it was proved that the compensator could make the PMSM asymptotically stable near the equilibrium point.Based on the modified interconnection and damping control,the Hamiltonian energy function was changed according to the expected equilibrium point,and a tracking controller with undetermined parameters was proposed.Then the GWO was applied to optimize the adjustable gains and undetermined parameters.Finally,the experiments have proved that the PMSM chaos is better suppressed,the system has greater following performance and ability to resist load disturbance.Given the benefits of PMSM chaos to the system and the universal significance of the chaotic synchronization control,a RBF-GWO chaotic synchronous controller with GWO and various variables was designed according to the radial basis function(RBF)neural network.The GWO was applied to optimize the parameters of RBF network,the central matrix,the output weight and the width vector with the minimum synchronization mean square error,hence the designed RBF-GWO network has the optimal performance.The effectiveness of the proposed chaotic synchronous controller was verified from the aspects of chaotic homogeneous and heterogeneous synchronization of PMSM,and the chaos phenomenon of PMSM could be utilized more rationally.On Par4 parallel robot,the high-speed pick-up trajectory planning and tracking control was carried out.For the high-speed pick-up path of Par4 parallel robot,a trajectory planning method based on the Lamé curve was proposed,in which the fifth and sixth asymmetric polynomials has been taken as the motion law,and the GWO was adopted to optimize the trajectory with the purpose of minimizing the mechanical energy consumption of the robot.Finally the trajectory with the lowest energy consumption was found,which verified the effectiveness of the planning method.In addition,it also reveals that the parameter e optimal value of the Lamé curve could be selected as half of the pick-up span,and the optimal value of the parameter f should be set according to the pick-up coordinate and the pick-up height.Taking the planned motor angles of Par4 parallel robot as the expected inputs,a trajectory tracking PID controller based on Type2-fuzzy forecast compensation was designed,in which the sum of the input change ratio and the tracking error change ratio was as an input of Type2-fuzzy forecast compensation,which has improved the dynamic following performance of the system and reduced the angle tracking error of the driven motor.The dynamic GWO was adopted to optimize the Type2-fuzzy controller off-line to boost the performance of the system.Finally,the effectiveness of the proposed controller was verified by experiments.The four drive motors of Par4 parallel robot could better track the desired angle input.