Application Of Partical Swarm Optimization And Differential Evolution Based-on Hierarchical Muti-subpopulation In Inverse Kinematics

Giving the location or coordinate of the robot end executor to solve the joint variable is the inverse kinematics of the robot, the inverse kinematics problem solving occupies a significant position in robotics, which is the foundation of the robot control.According to the robot kinematics equation established by D-H method, the robot joint angle and spatial coordinates of the end executor present serious nonlinear relationship.To the robot meeting the specific structure, in the case of a given end executor coordinates, we can continuously use parsing ends of the forward kinematics equation to inversely multiply transformation matrix through analytic method, to separate the joint variables and exact solutions will gradually be obtained.For the robots which does not meet the specific structure, analytic method can’t be used. Aiming at this problem, this paper uses partical swarm optimization and differential dvolution based-on hierarchical muti-subpopulation, by using conventional Evolutionary Algorithm, to constantly search the optimal solution in the solution space of joint variables which meet the conditions with the robot kinematics equation.Firstly, this thesis elaborates the relations between robot kinematics equation transformation and the connecting rod in the coordinate system, and on this basis, deduces two kinds of typical robot:spray PUMA560 robot and the forward kinematics equation of spray robot. Secondly, some improvements are made on the basis of the standard particle swarm optimization algorithm(PSO) and differential evolution algorithm, adaptive inertia weight and learning factor in the improved particle swarm optimization algorithm make corresponding changes during iterative process, which can better adapt to the evolution of the whole population. At the same time, the original mutation operator in difference algorithm will be replaced by adaptive mutation operator, which makes the algorithm has better convergence properties.Then improved particle swarm optimization and differential algorithm will be fused into partical swarm optimization and differential dvolution based-on hierarchical muti-subpopulation, multiple populations at the bottom of the adaptive particle swarm optimization algorithm is adopted to do coarse search, the completions of each population form the top part after search and then do fine search with the differential evolution algorithm. Multiple populations of hierarchical topology structure can sort the performance of the particle population from bottom to top, from low to high, is therefore more able to balance the global search ability and local search ability. The simulation of classical test functions show that the fused algorithm has great improvement than particle swarm optimization and differential algorithm in convergence speed and precision of calculation. Finally, the thesis uses the proposed algorithm in PUMA560 robot and spraying robot, do the inverse kinematics solving and simulation in MATLAB experiment platform, the experimental results show that the algorithm in this paper is an effective alternative when we cannot directly solve or solve relatively complex situations using the analytic method.