This paper presents Autoware on Board, a new profile of Autoware, especially designed to enable autonomous vehicles with embedded systems. Autoware is a popular open-source software project that provides a complete set of self-driving modules, including localization, detection, prediction, planning, and control. We customize and extend the software stack of Autoware to accommodate embedded computing capabilities. In particular, we use DRIVE PX2 as a reference computing platform, which is manufactured by NVIDIA Corporation for development of autonomous vehicles, and evaluate the performance of Autoware on ARM-based embedded processing cores and Tegra-based embedded graphics processing units (GPUs). Given that low-power CPUs are often preferred over high-performance GPUs, from the functional safety point of view, this paper focuses on the application of Autoware on ARM cores rather than Tegra ones. However, some Autoware modules still need to be executed on the Tegra cores to achieve load balancing and real-time processing. The experimental results show that the execution latency imposed on the DRIVE PX2 platform is capped at about three times as much as that on a high-end laptop computer. We believe that this observed computing performance is even acceptable for real-world production of autonomous vehicles in certain scenarios.
Kato et al. (2018) studied this question.