复杂多关节机械臂建模及逆运动学比较分析

针对含多种关节类型的复杂机械臂的建模问题, 在 D鄄H 参数法中引入虚拟关节建立运动学模型。 考虑机械臂逆运动学存在多解、 精度和实时性的问题, 通过4 种不同的 PSO(Particle Swarm Optimization)优化算法: 线性递减权重的粒子群(LPSO: Linear Decreasing Weight Particle Swarm Optimization)、 基于杂交的粒子群优化(CBPSO: Crossbreed Particle Swarm Optimization)、 基于模拟退火的粒子群(SAPSO: Simulated Annealing Particle Swarm Optimization)和混沌粒子群优化(CPSO: Choas Particle Swarm Optimization)进行计算。 随机选取工作空间的位置点, 验证优化算法能有效计算机械臂逆运动学解, 并对执行时间、 位置误差等方面进行了比较分析。 实验结果表明, 改进的 CBPSO 算法能有效计算复杂多关节机械臂的逆运动学解, 同时满足实际作业中对实时控制的要求。

Comparison and Analysis of Kinematic Model and Inverse Kinematics of Complex Multi-Joint Manipulator

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.