RBF networks-based inverse kinematics of 6R manipulator

From the point of view of set theory and mathematics, the relation between the forward kinematics (FK) and the inverse kinematics (IK) can be regarded as a nonlinear mapping between the joint space and the operation space of the robot manipulator. Considering the powerful ability of the artificial neural networks (ANN) to process nonlinear mapping relations, the IK problem can be transformed into the problem of training the weights of ANN. In this work, the solution of the IK of the MOTOMAN manipulator is implemented by using ANN. Because of its local approach ability, the radial basis function (RBF) networks of six inputs and one output are designed. The method avoids the traditional complicated deriving equations procedure and programming. Examples are given to illustrate that RBF networks not only have better computation precision than back propagation (BP) networks but also converge faster than BP networks.