Solving Inverse Kinematics of Industrial Robot Based on BP Neural Network

摘要

Inverse kinematics of robot means solving the joint variables by the position and posture of robot end-effector, which is the basis of robot control. For the traditional method of inverse kinematics is more complex and not universally applicable, an algorithm based on BP neural network solving inverse kinematics of robot is proposed in this paper. Taking RS10N industrial Robot as the research object, a BP neural network is trained and improved to solve its inverse kinematics, which is simulated by MATLAB toolbox. The simulation results show that BP neural network trained by LM algorithm is precise and fast, which is a reliable method of solving inverse kinematics of robot.