COOPERATIVE STATE ESTIMATION AND MOTION COORDINATION OF MULTIPLE AUTONOMOUS VEHICLES

Abstract

Autonomous driving is challenging due to limited perception range, sensing occlusion, unpredictable future trajectories of other vehicles, and unforeseeable motion conflicts between vehicles. These challenges can be overcome much more easily with vehicle-to-vehicle (V2V) communication, which allows information sharing among vehicles. In order to tackle these challenges, we study the problem of multi-vehicle cooperative state estimation and motion coordination, where V2V communication is adopted. Specifically, we introduce the minimal sensor configuration for scalable cooperative localization, propose a general framework of cooperative localization to account for communication delay and loss, perform impact analysis of communication loss on cooperative state estimation, and propose our motion coordination algorithm. The results can help autonomous vehicles to reduce sensor cost, increase localization accuracy for information fusion, and improve driving safety and efficiency.