Robot arm reaching through neural inversions and reinforcement learning

We present a neural method that computes the inverse kinematics of any kind of robot manipulators; both redundant and non-redundant. Inverse kinematic s solutions are obtained through the inversion of a neural network that has been previously trained to approximate the manipulator forward kinematics. The inversion provides difference vectors in the joint space from differen ce vectors in the workspace. Our differential inverse kinematics (DIV) appr oach can be viewed as a neural network implementation of the Jacobian trans pose method for arm kinematic control that does not require previous knowle dge of the arm forward kinematics. Redundancy can be exploited to obtain a special inverse kinematic solution that meets a particular constraint (e.g. joint limit avoidance) by inverting an additional neural network The usefu lness of our DIV approach is further illustrated with sensor-based multilin k manipulators that learn collision-free reaching motions in unknown enviro nments. For this task, the neural controller has two modules: a reinforceme nt-based action generator (AG) and a DIV module that computes goal vectors in the joint space. The actions given by the AG are interpreted with regard to those goal vectors. (C) 2000 Published by Elsevier Science B.V. All rig hts reserved.