Improvement in Localization of a Moving Vehicle using K-means Clustering

Abstract

Automotive radar uses multiple sensors to locate and determine the movement of the vehicles. Frequency Modulated Continuous Wave (FMCW) is the prevailing waveform used in this field. The proposed method estimates the vehicle location and speed. It tracks the vehicle(s) using multiple sensors arranged in a Uniform Linear Array (ULA). Vehicle identification and assigning tracks to a vehicle in case of a multi-target scenario is based on three estimated parameters: vehicle range, speed, and range-to-track association. Both single and multiple target scenarios are examined in the proposed method. The study is further extended to a uni-directional multi-lane road scenario to establish the universality of the method.