V2V 기반 차선 검출시스템 개발

Abstract

This paper proposes the virtual lane detection system in order to classify the surrounding vehicles. The virtual lanes are generated by analyzing the movements of the surrounding vehicles which are received from V2V (Vehicle-to-Vehicle) communication. Many active safety systems using radar and camera have blind spot zones caused by the FOV(Field of View) or the surrounding structure. The proposed system uses V2V communication in order to classify the relative positions of the surrounding vehicles in blind spot zones. Furthermore, by using the path history of vehicles, more accurate target classification becomes available for active safety systems such as FCW (forward collision warning), BSW (Blind Spot Warning), IMA (Intersection Movement Assist) and DNPW (Do Not Pass Warning). The commercial software was performed to validate the proposed system.