机器人运动轨迹神经网络模糊控制优化

Neural Network Fuzzy Control Optimization of Robot Trajectory

为了降低机器人关节运动轨迹误差,设计了神经模糊控制系统,通过仿真检验机器人关节跟踪误差。创建机器人关节结构简图,给出机器人动力学方程式,引用三阶多项式来保证机器人关节运动轨迹连续性;结合神经网络结构和模糊规则,设计机器人神经网络模糊控制器,利用遗传算法原理对神经网络模糊控制器进行优化;采用MATLAB软件对机器人关节运动角位移进行仿真,并且与优化前进行比较。结果表明,在空载环境中,优化前与优化后的神经网络模糊控制系统,机器人关节角位移输出误差相对较小,差别不大。在负载环境中,优化前的神经网络模糊控制系统,机器人关节角位移输出误差较大,优化后的神经网络模糊控制系统,机器人关节角位移输出误差较小。采用遗传算法优化神经网络模糊控制系统,能够有效提高机器人关节运动轨迹输出精度。

In order to reduce the robot joint trajectory error,a neuro fuzzy control system is designed,and the robot joint tracking error is verified by simulation.The structure diagram of robot joint is created,the dynamic equation of robot is given,and the third-order polynomial is used to ensure the continuity of robot joint trajectory;Combined with the neural network structure and fuzzy rules,the robot neural network fuzzy controller is designed,and the neural network fuzzy controller is optimized by using the principle of genetic algorithm;The angular displacement of robot joint motion is simulated by MATLAB software,and compared with that before optimization.The results show that in the no-load environment,the output error of robot joint angular displacement is relatively small and there is little difference between the optimized neural network fuzzy control system and the optimized neural network fuzzy control system.In the load environment,the output error of robot joint angular displacement is large in the optimized neural network fuzzy control system,and the output error of robot joint angular displacement is small in the optimized neural network fuzzy control system.The neural network fuzzy control system optimized by genetic algorithm can effectively improve the output accuracy of robot joint trajectory.