Neural network learning from hint for the cyclic motion of the constrained redundant arm

A Widrow-Hoff neural network (NN) with an online adaptive learning algorithm derived by applying Lyapunov approach is introduced for the kinematic inversion of redundant arms. The developed approach is designed to enable the manipulator to conserve the joint configuration in cyclic trajectories and to avoid the joint limits. Since the inverse kinematics in this problem has an infinite number of joint angle vectors, a fuzzy neural network (FNN) is designed to provide an approximate value for that vector. Feeding this vector as an additional hint input vector to the NN limits and guides the output of the NN within the self-motion. The derivation of the Lyapunov function which is designed to achieve both of the tasks, leads to a computationally efficient online learning algorithm of the NN. The effectiveness of the developed approach is studied by conducting experiments on the PA-10 redundant manipulator