The main research directions of the lab include fusion perception and intelligent decision-making. China has the most complex mixed traffic scenes in the world, and the accurate and reliable perception is the key bottleneck technology for breakthroughs in high-level autonomous driving. The lab focuses on building a vehicle-road-cloud multi-modal perception fusion theory with maps as a super perception container, establishing a Chinese standard autonomous driving AI perception system, and realizing fusion integration and accurate perception and cognition in complex traffic scenarios in China. At the same time, on the basis of perception and cognition, explore the safety decision-making mechanism for autonomous driving, study online learning theories such as deep reinforcement learning, and realize the self-evolution and migration of decision-making ability for autonomous driving.