机器人逆运动学的模拟退火自适应遗传算法

Adaptive Genetic Algorithms Based on Simulated Annealing for the Inverse Kinematics of the Manipulator

对5自由度关节型机器人的运动特点进行分析,并建立该机构的运动模型.在基本遗传算法的基础上,针对求解串联机器人运动学逆解的特殊性,为了消除遗传算法易陷入局部最优或早熟问题,提出一种新颖的改进遗传算法,采用模拟退火和自适应策略,保证其收敛性并加快收敛速度.仿真结果表明,该算法求解精度高,收敛速度快,且稳定性好.

The kinematics model of 5R manipulator is constructed on the basis of the analysis of its kinematics characteristic. A novel modified genetic algorithm is presented after consideration of particularity of inverse kinematics problem for serial robot and elimination of simple genetic algorithm＇s getting into local optimization or premature convergence. Simulated annealing and adaptive strategy are applied in the algorithm for its convergence process. The simulation showed that the algorithm is characterized by high precision, quick convergence and stable property.