Automotive Radar – Road Boundary Estimation

Abstract

The automotive industry is progressively focusing on autonomous driving systems, and there are two main reasons behind that: first, the amount of legislation related to autonomous vehicles is gradually increasing, and second, the OEMs continuously seek to enhance the vehicles safety and comfort. Autonomous Driving Assistant Systems use multiple sensors to sense the 360 degrees all around a vehicle in order to mitigate any risk of collision. One of the key sensors in this field is the radar. This sensor is usually mounted as a front radar to cover the driving direction, or as a corner radar. A radar system can deliver information about stationary or moving targets to a fusion system. The typical interface of radar system is either a user warning alert or an object with properties such as its position, velocity, orientation, its qualities and other parameters. The higher the automation level, the greater the amount of information required by the system. The end user warning functions and tracked objects are a standard kind of information over the years. However, carmakers are currently focusing on low-level fusion and hence require more information and features from the sensors. A typical use case is to deliver the road boundaries over rideable obstacles. Therefore, the aim of this paper is to introduce the guardrail detection techniques with focus on the measurement of a curbstone using the 4D technology.