Development of a Robot Car by Single Line Search Method for White Line Detection with FPGA

Abstract

In level 5 autonomous driving system, image recognition is required as multiplex safety technology. However, real time image recognition is hard for existing microprocessors. Hence, implementation of driving system on FPGA is useful to achieve real time image recognition for autonomous driving. Therefore, this paper describes implementation of autonomous driving robot with image processing using FPGA. Hough transform which is generally used for white line detection, requires high computing cost. We explain our new white line detection method which features low computation cost. As a result of evaluation, image processing performance on software is about 26.1 frame / sec, and on hardware is about 0.5 frame / sec. In addition, hardware implementation using Vivado HLS is described.