Parallel Driving OS: A Ubiquitous Operating System for Autonomous Driving in CPSS

Abstract

With the rapid development of autonomous driving technologies, a vast array of autonomous driving algorithms and platforms have emerged. These algorithms and platforms are usually designed for different tasks in specific scenarios, which makes them incompatible with each other. To solve the problem, we propose a ubiquitous operating system named Parallel Driving Operating System (PDOS) based on the parallel driving theory. Specifically, the PDOS architecture includes the hardware layer, the kernel layer, the functional layer, and the application layer, which can realize heterogeneous hardware support, ubiquitous resource management, algorithm optimization, and ubiquitous application development, respectively. Moreover, we derive the PDOS into specific operating systems in four application scenarios, i.e., intelligent mining, intelligent warehousing, intelligent logistics, and intelligent port. To validate the PDOS, we test its usability, reliability, and efficiency in these scenarios. Evaluation results demonstrate the high performance of PDOS in multiple autonomous driving, and indicate that PDOS is an efficient operating system for parallel driving.