Robust, Adaptive Control of Systems with Nonlinearities

Weapon system components such as electromechanical or hydraulic servos, valves, or gears often exhibit deadzone, backlash, or hysteresis behavior caused by stiction, magnetism, or other phenomena. These nonlinear features can strongly influence or dominate system response in many situations. Feedback control laws for such systems must manage the impact of these types of nonlinearities on system performance. However, it is frequently difficult to characterize these nonlinear effects precisely, and this hinders the design of effective control laws. In this proposal we describe a design methodology for the robust, adaptive feedback control of systems with backlash, hystereses, and deadzones. We shall investigate the tradeoffs in complexity between classical robust control methods (Hinfinity control) and parametric adaptive control methods. The adaptive control methodology is based on the use of (adaptive) inverses for the nonlinear phenomena (backlash, hysteresis, and deadzones) and adaptive update laws for the linear plant uncertainties. The methods have produced good results for small problems. In this work we will (i) extend these adaptive control techniques to treat more complex, MIMO systems; (ii) compare their performance with H-infinity control designs for a test article resident at the Picatinny Arsenal.