Iterative Learning of Impedance Parameters for Manipulator Control Using Neural Networks

Abstract

Impedance Control is one of the most effective control methods for the manipulators in contact with their environments. The characteristics of force and motion control, however, are determined by the desired impedance of end-effectors which must be designed according to the given tasks. In the present paper, we propose a method to regulate impedance parameters of the manipulator's end-effector using neural networks. Three kinds of the Back Propagation typed neural networks are prepared corresponding to position, velocity and force control of the end-effector. Firstly, the neural networks for position and velocity control are trained using iterative learning of the manipulator during free movements. Then, the neural network for force control is trained for contact movements. During the contact movement, the virtual trajectory of the end-effector is modified to reduce the force control error. Computer simulations shows that the method can regulate not only stiffness and viscosity but also inertia of the end-effector, and can realize smooth transition from free to contact movements by regulating impedance parameters before contact.