Modification of DBSCAN and application to range/Doppler/DoA measurements for pedestrian recognition with an automotive radar system

Abstract

We present in this paper modifications of the Density-Based Spatial Clustering of Applications with Noise (DBSCAN) algorithm in order to detect pedestrians with a conventional automotive 77-GHz frequency modulated radar system. These modification include dimension scaling as preprocessing and a generalization of the ε-neighborhood notation by introducing a size parameter to constitute the new Ellipsoid DBSCAN (EDBSCAN) algorithm. With these modifications we could successfully cluster real-world measurement data in order to get reasonable cluster representations of pedestrians in a cluttered environment.