Distributed System Architecture of Autonomous Vehicles and Real-Time Path Planning Based on the Curvilinear Coordinate System

The development of autonomous vehicle requires the state-of-the-art technologies in perception, planning, control, and system integration. This paper presents an overview of the system architecture and software architecture of autonomous vehicles for system integration. Network based system architecture in this paper provides a distributed computing system for autonomous driving. Further, a real-time path planning and a target speed generation are described based on the curvilinear coordinate system. The design of a path in the curvilinear coordinate system stretches the design space as like the Cartesian coordinate system to simplify the generation of the path. In determination of target speed, curvatures and risk of a generated path were utilized for safe autonomous driving. The proposed system architecture and planning algorithm were successfully integrated into the autonomous vehicle A1, which was developed by the Automotive Control and Electronics Laboratory (ACE Lab) and Machine Monitoring and Control Laboratory (MMC Lab) from Hanyang University, Seoul, Korea.