Networking and Communications in Autonomous Driving: A Survey

The development of light detection and ranging, Radar, camera, and other advanced sensor technologies inaugurated a new era in autonomous driving. However, due to the intrinsic limitations of these sensors, autonomous vehicles are prone to making erroneous decisions and causing serious disasters. At this point, networking and communication technologies can greatly make up for sensor deficiencies, and are more reliable, feasible and efficient to promote the information interaction, thereby improving autonomous vehicle's perception and planning capabilities as well as realizing better vehicle control. This paper surveys the networking and communication technologies in autonomous driving from two aspects: intra- and inter-vehicle. The intra-vehicle network as the basis of realizing autonomous driving connects the on-board electronic parts. The inter-vehicle network is the medium for interaction between vehicles and outside information. In addition, we present the new trends of communication technologies in autonomous driving, as well as investigate the current mainstream verification methods and emphasize the challenges and open issues of networking and communications in autonomous driving.

I. Introduction

In recent years, both the academia and industry have shown great interest in the development of autonomous driving, which will liberate drivers physically and mentally, greatly improve traffic safety and energy efficiency, as well as make better use of public resources. Universities and research groups are actively involved in autonomous driving competitions and technical challenges. Stanford University and Carnegie Mellon University have done well in “DARPA Urban Challenge” [1], a top class competition for autonomous driving techniques organized by DARPA (Defense Advanced Research Projects Agency). Traditional car companies such as General Motors, Toyota, Volvo, Volkswagen and BMW as well as Internet auto companies such as Google, Uber and Tesla, have joined the ranks of autonomous vehicle manufacturing and research. SAE International divides autonomous driving into 5 different levels [2]. Tesla Model S claims to have reached level 2.5 of autonomous driving; the new Audi A-8 has reached level 3, making Audi the first automaker to offer a level 3 autopilot. Although the final winner in this area is still difficult to judge, the emergence of autonomous vehicles will make a historic change in the way people and goods are transported.