Template matching for radar-based orientation and position estimation in automotive scenarios

Abstract

Future high-resolution radars enable new functionalities in advanced driver assistance systems, such as estimation of contour, position, and orientation of vehicles on the road. However, straightforward approaches like that of Oriented Bounding Box generally fail in challenging automotive scenarios, when only one side of the vehicle is visible to the radar. In this paper, an estimation approach based on Generalized Hough Transform matching is presented and examined for its use in such automotive scenarios. An optimization method is discussed and finally, the algorithm is compared against a classic approach.