Economic MPC With Adaptive Sampling and Offset-Free Operation for Perturbed Autonomous Electric Vehicles

This article addresses the problem of energy-efficient driving of autonomous electric vehicles with economic performance and desired tracking effect. To handle the issue, an economic model predictive control (EMPC) algorithm with adaptive sampling and offset-free operation is proposed by designing self-triggered mechanisms with consideration of economic criteria and state error estimation. We synthesize an observer with an offset-free setup and an economic model predictive controller in the proposed algorithm. Under an offset-free design, the observer estimates the actual state by introducing a compensation item to a prediction model, and thus a bias between an actual state and a predicted state is eliminated. An upper bound that limits the predicted error to a tightened set is derived by constraint tightening theory. In addition, the optimal intersampling time required to transmit optimal control input is calculated by two self-triggered mechanisms. Switching between these two self-triggered mechanisms can further maximize the optimal intersampling time and reduce the computational complexity. Recursive feasibility and asymptotic average performance are ensured for the tracking error system. We prove that the state trajectory of the system eventually converges to optimal steady-state under self-triggered mechanisms. The efficacy and advantages of the proposed algorithm are verified by simulation and experimental results.