Inverse kinematic in SCARA manipulator based on RBF networks

In this article we have provided a method of neural networks based on radial basis functions which is also named solving method of RBF (Radial Basis Function) neural network to deal with inverse kinematic. This method used some forward kinematic results of SCARA robot which Harbin Kelida Co Ltd provided as training samples, and mapping from workspace to joint space was obtained with iterative training, then calculation of inverse kinematic was achieved. It avoided complex process of formula deduction and programming calculation, its derivation was simple and its character of real time was good.