4C: A Computation, Communication, and Control Co-Design Framework for CAVs

Connected and autonomous vehicles (CAVs) are promising due to their potential safety and efficiency benefits, and have attracted massive investment and interest from government agencies, industry, and academia. With more computing and communication resources available, both vehicles and edge servers are equipped with a set of cam-era-based vision sensors, also known as Visual IoT (VIoT) techniques, for sensing and perception. Tremendous efforts have been made to achieve programmable communication, computation, and control. However, they are conducted mainly in silo mode, limiting the responsiveness and efficiency of handling challenging scenarios in the real world. To improve the end-to-end performance, we envision that future CAVs require co-design of communication, computation, and control. This article presents our vision of the end-to-end design principle for CAVs, called 4C, which extends the VIoT system by providing a unified communication, computation, and control co-design framework. With programmable communications, fine-grained heterogeneous computation, and efficient vehicle controls in 4C, CAVs can handle critical scenarios and achieve energy-efficient autonomous driving. Finally, we present several challenges to achieving the vision of the 4C framework.

Introduction

The recent proliferation of computing technologies (e.g., sensors, computer vision, machine learning, and hardware acceleration) has promoted connected and autonomous vehicles (CAVs) [1]. The global CAV market is expected to grow to $173.15 billion by 2030, with shared mobility services contributing 65.31 percent [2]. In general, there are three main benefits of CAVs. The first bene-fit is safety. According to the fatality analysis from the National Highway Traffic Safety Administration (NHTSA), 94 percent of serious crashes are caused by human error [3]. Compared to a human driver, the machine does not experience fatigue, or be drunk or speeding. The sensors are better at distance detection, blind space detection, emergency obstacle avoidance, and so on. The second benefit is efficiency for many aspects: less traffic congestion, less fuel consumption, less greenhouse gas emission, less travel time, and so on. The improvement is because CAVs get more comprehensive traffic information, and have better planning and controls than human drivers. The third benefit is that CAVs could support third-party applications, including applications for public safety like AMBER Alerts and criminal face detection. All these applications leverage the on-vehicle computation and communication resources.