Abstract: o establish kinematics model D-H parameter method is employed and virtual joints is introduced, according to a variety of joint types of complex manipulator. There are many problems of multi-joint manipulator, such as multiple solution, accuracy and real-time performance. The solution of the inverse kinematics is obtained by using four different optimization algorithm: the particle swarm optimization algorithm with LPSO(Linear Decreasing Weight Particle Swarm Optimization), CBPSO (Crossbreed Particle Swarm Optimization) algorithm, SAPSO (Simulated Annealing Particle Swarm Optimization) algorithm and CPSO(Choas Particle Swarm Optimization) algorithm. These algorithms are tested for the validity of the solution of inverse kinematics. Through randomly selected workspace points, the four algorithms are compared according to the executing time, the end-effector position error and so on. The results show that the improved CBPSO algorithm can be effectively used for the inverse solution for complex multi-joint manipulator, and satisfy the requirement of real-time control in actual work.

Key words: kinematics model, particle swarm optimization(PSO) algorithm, inverse kinematic,  multi-joint manipulator