Comparison of Event-Triggered Model Predictive Control for Autonomous Vehicle Path Tracking
Jun Chen, Zonggen Yi
Abstract
This paper proposes two different event-triggered nonlinear model predictive controls (NMPC) for autonomous vehicle path tracking. The difference between the two event-triggered NMPCs is the determination of control action when an event is not triggered. In the first formulation, the optimal control sequence computed from last triggering event is shifted to determine control action when NMPC is not triggered, while in the second formulation, a time-triggered linear parametric varying MPC (LPV-MPC) with shorter prediction horizon is formulated and solved in between NMPC triggering events to compensate prediction error and disturbance. These two event-triggered NMPCs, together with a time-triggered LPVMPC and a time-triggered NMPC serving as benchmark, are implemented to track the vehicle path in both longitudinal and lateral directions, with axle driving torque and front steering input as the control variables. Control performance and throughput requirements of different MPCs are then measured and compared, where the advantage of event-triggered formulation is clearly demonstrated.