Adaptive optimal control for the uncertain driving habit problem in adaptive cruise control system

摘要

In this paper, a novel adaptive optimal control approach based on Q-function is proposed to address the problem as the driving habits change among drivers and over time in the adaptive cruise control system. The proposed approach, adopting the special structure Q-function of the linear discrete-time system, uses policy iteration method to derive the optimal control policy online. It repeats between policy evaluation where the polynomial neural network is employed to approximate the cost function of the system and policy improvement where the control policy is updated based on the converged neural network, until the optimal controller is achieved. Simulation is conducted and results show the effectiveness for uncertain driving habit problem in the adaptive cruise control system.