Real-Time Error Compensation Strategy Based on BP Neural Network for Master-Slave Control

In order to deal with the nonlinearity of robot kinematics and the difficulty in solving inverse kinematics, with the help of Jacobian matrix and linearization thought, the mapping between small increment in Cartesian space and joint space is established. To eliminate the non-linear error due to ignoring high-order items, a real-time error compensation strategy based on neural network is proposed. The simulation based on 3-degree of freedom (3-DOF)master-slave system is carried out to verify the proposed method.