Integration of vehicle and lane detection for forward collision warning system

Abstract

The capability of automated vehicles with intelligent on-board systems aimed at improving traffic safety has been in constant advance. This paper proposes an integrated module of lane and vehicle detection for a forward collision warning system that can be embedded in an autonomous driving system operating in real time. Integration of lane and vehicle detection provides more precise information for driving environments by using geometric consideration of the roads, and achieves synergistic effects for rejecting false alarms by adjusting ground constraints. Also, once the vehicle is localized in ego-lane, it is able to estimate the distance between the front and ego-vehicle. The experimental results in this paper, conducted on a real road dataset, verify the applicability of the proposed work for advanced driver assistance systems.