Application of RBF Neural Network in Trajectory Planning of Robot

Abstract

Trajectory planning of robot is to control the robot in order to accurately follow the target track. And the target trajectory is always high-order and nonlinear. But RBF neural network can be achieved from the input to the output of arbitrary nonlinear mapping, through network learning and training to achieve the nonlinear function. This paper establishes a RBF neural network model firstly, and carries on the simulation through software MATLAB. The result confirms that the RBF neural network can keep the control of the robot's nonlinear trajectory planning in real time.