Heterogeneous Multiple Vehicles Cooperation Approach for Smart Roads

Intelligent vehicles are equipped with multiple on-board sensors for environment perception. Moreover, with the increasing number of these vehicles on the roads, the more cooperation and coordination among them is becoming more crucial. Accordingly, this paper presents multiple heterogeneous vehicles cooperation approaches, to be used in smart roads to improve driving safety. The heterogeneity aspect is based on the use of Unmanned Aerial Vehicle (UAV) to scout the surrounding of the Unmanned Ground Vehicle (UGV), thus increasing the perception efficiency. Two approaches were proposed for this cooperation, vehicle and pedestrian detection. The algorithms are implemented in the on-board computers. In order to evaluate the proposed approaches, different scenarios were selected and multiple experiments were carried out. The obtained results show the high performance of the algorithm in almost real-time detection and classification, moreover the ability to communicate the outcomes to the UGV, thus improving the automated navigation process for out of the line of sight pedestrians.