Abstract
In this paper, a cooperative vehicle control system to avoid collision with an obstacle is proposed. The model predictive control (MPC) combined with environment description by potential field (PF) function is suitable for such a system to generate a target trajectory because of its generality of problem formulation and real-time re-planning capability to deal with situation change. To make the MPC-based control algorithm cooperative with a human driver, two techniques are proposed. One is a direction selective trajectory generation to match human driver’s intention and a control strategy of the system. The other is an appropriate control command shaping considering human driver’s response against assist control from the system to prevent interference between the human driver and the system. Simulation results show each proposing technique works effectively.
Keywords: Intelligent Vehicles, Cooperative Driver Assistance Systems, Model Predictive Control, Potential Field.