Vehicle location estimation based on 79GHz UWB radar employing road objects

Abstract

This paper presents a vehicle location estimation based on 79GHz ultra-wideband (UWB) radar, which utilizes various road objects around car. Lane-keeping-assist system with camera is available in many cars to keep self-vehicle inside a solid white line, which is one of important systems for self-driving car. However, camera may be difficult to recognize the line in bad weather conditions, while 79GHz UWB radar may recognize the line with ribs. In this paper, the effect of snowfall on the detection performance has been investigated. Next, vehicle location estimation scheme at 79GHz UWB radar is suggested instead of white line where the measurement results of roadside objects such as guardrail, bridge-rail, and pole-corn instead of the line are presented.