Intelligent compensation of actuator nonlinearities

Inaccuracy of mechanical components and nature of physical laws make actuators nonlinear devices. Actuator nonlinearities can be static like friction, deadzone, saturation, and/or dynamic like backlash and hysteresis. In this paper, we present recent developments in four basic problems in actuator control-friction, deadzone, backlash, and saturation. We apply intelligent control techniques for compensation of those nonlinearities for a class of nonlinear systems. Static and dynamic compensators using neural networks are considered. Different control structures are developed for common actuator nonlinearities. We present tuning algorithms for those intelligent compensation techniques that guarantee small tracking error and bounded internal states. Simulation results show that the proposed compensation schemes are efficient way of improving the tracking performance of nonlinear systems with backlash.