重载精密机械手末端误差的多目标优化
Multi-objective Optimization of Error in Heavy-load Manipulator

针对关节几何误差对重载机械手运动精度的影响，鉴于PID神经元具有多输入和多输出优化的特性，提出重载精密机械手误差的多目标优化方法，研究重载机械手误差的关键因子，分析重载精密机械手的映射及机械手多输出目标之间的关系，探究PID控制器与神经元网络之间的对应关系，求解重载机械手精密运动的传递关系，同时推导机械手位姿、速度和加速度与关节几何误差之间的数学模型，得出机械手关节误差对其运动精度的影响趋势，以SCARA机器人为研究对象，采用PID神经元的多目标优化方法对其误差进行整合。结论表明：PID神经元的多目标优化方法有效地减小了关节几何误差对重载机械手的影响，其合理性得到了验证。
In view of the influence of joint geometric error on the motion and accuracy of a heavy-load manipulator, using PID neurons with multiple input and multiple output optimization characteristics, a multi-objective heavy-load manipulator error optimization method based on PID neurons is put forward. The key factor of precision manipulator's error is studied and the relationship between load mapping and manipulator's multi-precision output target is analyzed. Then, the relationship between PID controller and neural network and the transfer relationship for the heavy-load manipulator's movement precision are explored. The mathematical model between simultaneously derived manipulator pose, velocity and acceleration and joint geometry error is established. Finally, the trend of influence of the manipulator's joint error on the movement accuracy is studied and simulated with the SCARA software. The simulation results show that the influence of joint geometric errors on the heavy manipulator can be effectively reduced with the multi-objective optimization method based on the PID neural network, and the rationality of the method is verified