Decision Methodology Based on Dynamic Occupancy Grid Zone for Autonomous Lane Change System on Highways

Abstract

From the perspective of decision-making in autonomous lane change systems, it is very important for a vehicle to recognize itself a driving environment. In this paper, we propose a lane change decision methodology using full-range situation risk assessment based on a dynamic occupancy grid zone. The longitudinal length of the dynamic zone has flexible distances. To make the decision of lane change, collision risk assessment index (CRAI) and vehicle cooperative interaction index (VCII) are proposed to evaluate the driving situations. The CRAI is computed as a ratio of distance collected by automotive radar sensors and pre-defined distance of the dynamic zone. The VCII is available by analyzing an interaction with surrounding vehicles. To verify the methodology, scenario-based experiments were performed in real time with test vehicles in proving ground. In several experiments, the lane changes were performed successfully.