Implementation of a ROS-Based Autonomous Vehicle on an FPGA Board

Abstract

Due to the development of high-performance LSIs, autonomous vehicle will be realized in a few years. The FPGA Design Competition is one of the valuable opportunities to demonstrate an autonomous vehicle on miniature roads. In this paper, we develop a ROS-based autonomous vehicle which is implemented on an FPGA board as a mock car. ROS is a common framework designed to implement various types of robots. Utilizing the ROS-based platform, we develop a model car for the demonstration of an autonomous vehicle. The on-board programmable logic is used to off-load the processing of image recognition such as lane detection, traffic signal detection, and obstacle detection. The implementation results demonstrate that we can successfully implement essential components of the vehicle on an FPGA board with the ROS-based system.