基于遗传算法的护理机器人逆运动学求解方法

The nursing robot’s arm does not meet the PIEPER criterion and is constructed with a high degree of coupling and complexity. The standard genetic algorithm is difficult to accurately solve its inverse kinematics. As a result, the pose error of the end of the robot arm is relatively large. In order to solve the problem that premature and poor local search ability in the standard genetic algorithm solution process, the following methods are proposed. Firstly, use equal partitions to replace the initial population of individuals generated by random commands. Divide 5 small areas to improve the dispersion of the population and search efficiency. Secondly, introduce variable weight factors in the fitness function. The change value of the pose error is used as the variable of the variable weight factor. The weight factor varies between 0.5-1.0 during the evolution process. And effectively assign position and attitude error weights. Finally, verified by simulation and experiment. The results show that the improved genetic algorithm can greatly improve the accuracy and speed of convergence, achieve precise control of position and attitude at the same time, and greatly reduce the pose error of the robot arm.