Nasser L. Azad

Bijan Sakhdari, Nasser L. Azad

This paper proposes an adaptive tube-based nonlinear model predictive control (AT-NMPC) approach to the design of autonomous cruise control (ACC) systems. The proposed method utilizes two separate models to define the constrained receding horizon optimal control problem. A fixed nominal model is used to handle the problem constraints based on a robust tube-based approach. A separate adaptive model is used to define the objective function, which utilizes least square online parameter estimators for adaption. By having two separate models, this method takes into account uncertainties, modeling errors and delayed data in the design of the controller and guaranties robust constraint handling, while adapting to them to improve control performance. Furthermore, to be able implement the designed AT-NMPC in real-time, a Newton/GMRES fast solver is employed to solve the optimization problem. Simulations performed on a high-fidelity model of the baseline vehicle, the Toyota plug-in Prius, which is a plug-in hybrid electric vehicle (PHEV), show that the proposed controller is able to handle the defined constraints in the presence of uncertainty, while improving the energy cost of the trip. Moreover, the result of the hardware-in-loop experiment demonstrates the performance of the proposed controller in real time application.

Yuan Lin, John McPhee, Nasser L. Azad

The majority of current studies on autonomous vehicle control via deep reinforcement learning (DRL) utilize point-mass kinematic models, neglecting vehicle dynamics which includes acceleration delay and acceleration command dynamics. The acceleration delay, which results from sensing and actuation delays, results in delayed execution of the control inputs. The acceleration command dynamics dictates that the actual vehicle acceleration does not rise up to the desired command acceleration instantaneously due to dynamics. In this work, we investigate the feasibility of applying DRL controllers trained using vehicle kinematic models to more realistic driving control with vehicle dynamics. We consider a particular longitudinal car-following control, i.e., Adaptive Cruise Control (ACC), problem solved via DRL using a point-mass kinematic model. When such a controller is applied to car following with vehicle dynamics, we observe significantly degraded car-following performance. Therefore, we redesign the DRL framework to accommodate the acceleration delay and acceleration command dynamics by adding the delayed control inputs and the actual vehicle acceleration to the reinforcement learning environment state, respectively. The training results show that the redesigned DRL controller results in near-optimal control performance of car following with vehicle dynamics considered when compared with dynamic programming solutions.