Safety is the most important requirement for autonomous vehicles; hence, the ultimate challenge of designing an edge computing ecosystem for autonomous vehicles is to deliver enough computing power, redundancy, and security so as to guarantee the safety of autonomous vehicles. Specifically, autonomous driving systems are extremely complex; they tightly integrate many technologies, including sensing, localization, perception, decision making, as well as the smooth interactions with cloud platforms for high-definition (HD) map generation and data storage. These complexities impose numerous challenges for the design of autonomous driving edge computing systems. First, edge computing systems for autonomous driving need to process an enormous amount of data in real time, and often the incoming data from different sensors are highly heterogeneous. Since autonomous driving edge computing systems are mobile, they often have very strict energy consumption restrictions. Thus, it is imperative to deliver sufficient computing power with reasonable energy consumption, to guarantee the safety of autonomous vehicles, even at high speed. Second, in addition to the edge system design, vehicle-to-everything (V2X) provides redundancy for autonomous driving workloads and alleviates stringent performance and energy constraints on the edge side. With V2X, more research is required to define how vehicles cooperate with each other and the infrastructure. Last, safety cannot be guaranteed when security is compromised. Thus, protecting autonomous driving edge computing systems against attacks at different layers of the sensing and computing stack is of paramount concern. In this paper, we review state-of-the-art approaches in these areas as well as explore potential solutions to address these challenges.

To simultaneously enable multiple autonomous driving services on affordable embedded systems, we designed and implemented LoPECS, a Low-Power Edge Computing System for real-time autonomous robots and vehicles services. The contributions of this paper are three-fold: first, we developed a Heterogeneity-Aware Runtime Layer to fully utilize vehicle’s heterogeneous computing resources to fulfill the real-time requirement of autonomous driving applications; second, we developed a vehicle-edge Coordinator to dynamically offload vehicle tasks to edge cloudlet to further optimize user experience in the way of prolonged battery life; third, we successfully integrated these components into LoPECS system and implemented it on Nvidia Jetson TX1. To the best of our knowledge, this is the first complete edge computing system in a production autonomous vehicle. Our implementation on Nvidia Jetson demonstrated that it could successfully support multiple autonomous driving services with only 11 W of power consumption, and hence proves the effectiveness of the proposed LoPECS system.

Compressing massive LiDAR point clouds in real-time is critical to autonomous machines such as drones and self-driving cars. While most of the recent prior work has focused on compressing individual point cloud frames, this paper proposes a novel system that effectively compresses a sequence of point clouds. The idea to exploit both the spatial and temporal redundancies in a sequence of point cloud frames. We first identify a key frame in a point cloud sequence and spatially encode the key frame by iterative plane fitting. We then exploit the fact that consecutive point clouds have large overlaps in the physical space, and thus spatially encoded data can be (re-)used to encode the temporal stream. Temporal encoding by reusing spatial encoding data not only improves the compression rate, but also avoids redundant computations, which significantly improves the compression speed. Experiments show that our compression system achieves 40× to 90× compression rate, significantly higher than the MPEG’s LiDAR point cloud compression standard, while retaining high end-to-end application accuracies. Meanwhile, our compression system has a compression speed that matches the point cloud generation rate by today LiDARs and out-performs existing compression systems, enabling real-time point cloud transmission.

Autonomous driving is one of the most innovative applications nowadays. However, autonomous driving is still suffering from heavy calculation, high energy consumption and strict real-time execution constraints. Different from cloud computing, edge computing deploys calculation, storage and service on the edge of network. It is a better platform to serve efficiency and privacy oriented autonomous driving service offloading. To this end, we proposed a container-based edge offloading framework for autonomous driving. This framework builds an Offloading Decision Module, an Offloading Scheduler Module and an Edge Offloading Middleware on top of the lightweight virtualization. It provides the abstraction and management of the execution environment in the granularity of containers on edge. Therefore, it enables the privacy preserve and resource isolation for autonomous driving execution constraints. Its utility preferable offloading schedule strategy formalized the multi-application multi-edge nodes mapping problem into a multiple multidimensional knapsack problem (MMKP) and gave a utility oriented greedy algorithm (GA) for real-time solving. The experimental results show that the proposed framework has high feasibility and isolation meanwhile can guarantee millisecond-level autonomous driving offloading on edge.

This article introduces a technical solution for autonomous last-mile delivery in challenging traffic conditions. The methodologies in each module of the company's autonomous vehicles are presented together with safety guarantee strategies.

Autonomous Driving (AD) has entered a period of rapid development in recent years. With the amount of sensors and control logics installed increasing tremendously to guarantee robustness, a big challenge is posed for AD middleware. Both the academia and the industry are eager for an investigation of the performance of middlewares in Autonomous Driving Vehicles (AVs). To fill this gap, we summarize typical communication scenarios of AVs and evaluate different communication mechanisms of three popular open-source middlewares comprehensively. Besides, we construct a benchmark pack named ComP which consists of a perception communication scenario and a group of real AD applications for researchers to assess middleware performance. Our findings provide useful guidelines for researchers and insightful optimization advice for designing middlewares.