基于神经网络PID的机械臂末端力/位置混合控制系统

Hybrid Force/Position Control System of Manipulator-end Based on Neural Network PID

机械臂末端执行器与工作环境发生接触时候,对机械臂力控精度提出了更高的要求。面对工作环境的不确定性和变化,传统PID(Proportion Integral Differential)控制在机械臂力控方面表现出精度低,位置调节速度缓慢等特性。提出神经网络与PID控制相结合的智能控制策略,提高机械臂力控精度,调节速度与抗干扰性。运用拉格朗日法建立具有普遍研究意义的二连杆机械臂动力学模型,并通过雅克比矩阵计算出作用在机械臂各个关节的力矩。机械臂位置控制使用传统PID控制来满足动态调节速度,机械臂末端力控策略采用神经网络PID智能控制。通过MATLAB仿真实验结果表明:使用神经网络PID控制算法的机械臂具有良好的轨迹跟踪和力跟踪效果,为机械臂人机互动智能控制研究提供一定的参考。

When the end effector of the manipulator contacts with the working environment,higher requirements are put forward for the force control accuracy of the manipulator.In the face of the uncertainty and change of working environment,the traditional PID(proportion integral differential)control has the characteristics of low precision and slow position adjustment speed.An intelligent control strategy combining neural network and PID control is proposed to improve the precision of force control,speed regulation and anti-interference.The dynamic model of a two link manipulator with universal research significance is established by using Lagrange method,and the Cartesian force acting on the end of the manipulator is mapped into the equivalent joint torque through Jacobian matrix.The traditional PID control is used for the position control of the manipulator to meet the dynamic speed regulation,and the neural network PID intelligent control is adopted for the force control strategy of the manipulator end.The simulation results of MATLAB show that the robot arm with neural network PID control algorithm has good track tracking and force tracking effect,which provides a certain reference for the research of intelligent control of human-machine interaction of manipulator.