Research on Predictive Lane-changing Algorithm using Model Predictive Control
Yokohama2006/F2006V047

Abstract

In this research, an algorithm that determines an appropriate driving target in a situation where a driver wants to or has to make a lane change has been proposed based on Model Predictive Control (MPC), which is a control method that explicitly predicts the future behavior of the system and determines the optimal control input. An experimental system has been developed and installed on test vehicles. The system incorporates a wireless communication device for transmitting and receiving vehicle positions obtained by the Global Positioning System (GPS), as well as various vehicle control signals via an in-vehicle network. Vehicle-to-vehicle communication enables each vehicle to detect all other nearby vehicles. The MPC controller is designed to calculate optimal lane-change behavior and to target a preceding vehicle to be followed, in serving as an upper-level controller for conventional Adaptive Cruise Control (ACC) and an automatic Lane-Keeping Control (LKC) system. Predictions made by the MPC controller are expected to contribute to quicker response in executing lane changes. To evaluate the effectiveness of the algorithm, the driving maneuvers of human drivers were compared with and without the system in a situation which involved overtaking a slower moving vehicle ahead that was initially hidden by the nearest preceding vehicle. The results of experiments conducted on a proving ground course show that the system made lane changes earlier than the human drivers did, due to its ability to predict closing on the hidden vehicle and to decide immediately to avoid it.

Add to basket