Research On Active Safety Based On Multi-Vehicle Collaborative Perception

With the development of social economic, the number of automobiles increased rapidly. Traffic accident has become an increasingly severe social problem. It has directly threatened the safety of people's lives and properties. Therefore, vehicle safety has become a generally concerned issue. Although passive safety system can alleviate the damage of traffic accidents, it cannot prevent the accidents. Active safety system can effective detect the potential danger so as to prevent the accidents. However, the perception range of sensor in Single vehicle is limited. So, how to apply multi-vehicle collaboration technology is of great significance and attracted many researchers'attention. Multi-vehicle collaborative perception based active safety technology has been studied in this paper.The main research work are as follows: Vehicle accurate localization is the basis of vehicle collaboration in transportation system. Aiming at vehicle localization problem, a high accuracy RTK-GPS based vehicle localization method has been studied. In urban environment, GPS signal is easily blocked by tall buildings, crossing bridges, tunnels and etc. For compensating GPS deficiencies, Extended Kalman Filter algorithm is used for GPS and Dead Reckoning data fusion so as to guarantee the accuracy of localization. At the same time, a laser radar based RTK-GPS dynamic performance evaluation approach is proposed. The experimental results demonstrate that localization approach has high accuracy.Then, focusing on vehicle detection issue, a laser sensor based vehicle detection method is described. It applied corner point detection and template matching method for vehicle detection. On this basis, a multi-vehicle collaboration based vehicle detection method is presented. The experimental results show that multi-vehicle collaboration based vehicle detection method can greatly improve the environmental perception capability of vehicle.At last, considering traffic scenes of vehicle collision and vehicle lane-changing, the multi-vehicle collaborative pre-warning technology has been applied for preventing potential danger. Experimental results prove that multi-vehicle collaborative perception based active safety technology has brighten application prospect and can greatly enhance the vehicle safety in transportation system.