A Neural Network Approach for Inverse Kinematics Solving of Robots

This paper deals with manipulator inverse kinematics.A parameter disturbing algorithm of neural networks which can overcome the local minima porblem of Hopfield network is proposed.Its performance is analyzed in terms of digital simulations in detail.The algorithm is successfully applied to the inverse kinematics problem of manipulator control.It is shown by simulations that the proposed neural network controlling method not only has fast control speed but also raise the control accuracy of manipulator.