Advanced Driver Assistance Systems and Risk Identification in Cooperative Vehicle Infrastructure System Environment

Abstract

The existing advanced driver assistance systems (ADAS) can only detect targets in certain range while relying on on-board sensors, which cannot meet the requirements of collision warning at high speeds and long distances; Moreover, it is incapable for collision warnings at complex intersections and conflicts that may occur when the driver's sight is blocked. Under the framework of Cooperative Vehicle Infrastructure System (CVIS), the study proposed and realized the information interaction and targets fusion between vehicles and roadside devices. Based on the convenience of information interaction of CVIS technology, the on-board information interaction framework integrating vehicle and roadside sensing data is designed. In the proposed CVIS-based ADAS, on-board and roadside devices can separately detect obstacle vehicles according to their respective detection algorithms; The on-board system receives multi-source obstacle information and completes data fusion and map matching. On the basis of fused information in a unified coordinate system, the on-board system identifies vehicle collision risk and triggers vehicle collision warning. The implementation of this framework can expand the space-time application scope of risk identification and collision warning, and improve the adaptability of ADAS system in complex environment.