Design and Comparison of Explicit Model Predictive Controllers for an Electropneumatic Clutch Actuator Using On/Off Valves

This paper considers the optimal reference-tracking control of an electropneumatic clutch actuator for heavy duty trucks for which there exists an experimentally validated model. To control both supply to and exhaust from the clutch actuator chamber, one pair of on/ off valves is used. The fast dynamics of the actuator and the presence of state and input constraints makes it appropriate to design an explicit nonlinear model predictive controller (NMPC). In this paper, two different types of explicit NMPC controllers for optimal reference-tracking control of the electropneumatic clutch actuator are designed (one of the controllers applies a pulsewidth modulation (PWM) scheme, while the other controller allows the valves to be only fully opened or fully closed). The closed-loop performance of the two controllers is verified by simulating with the clutch actuator model. The controllers are compared with respect to control performance, real-time computational complexity, and storage requirements. A performance comparison with a sliding-mode controller and a PID controller shows that the explicit NMPC with PWM provides the highest reference-tracking quality. An appealing feature of explicit NMPC is that the performance improvement is feasible to implement with low computational complexity even for fast mechatronic systems.