Friction compensation of the electromechanical drive systems using neural networks

Friction is an undesired phenomenon in many mechanical systems. Feedforward of a suitable estimate of friction is an effective method to compensate for the friction-dependent position errors in the steady state. It is not easy, however, to make a precise friction model because of the complexity of static and dynamic characteristics of friction such as the Stribeck effect, the Dahl effect, stick-slip motion, and so on. In this paper, we propose an effective friction compensation method for the electromechanical drive systems. In the proposed method, neural networks are applied in parallel to a linear observer, for the electromechanical positioning system (EMPS) which is used at the Cologne Laboratory of Mechatronics (CLM) at the University of Applied Sciences Cologne for experimental investigation of position control schemes for compliant systems with friction.