Advanced Radar Signal Processing for Enhanced Object Detection in ADAS

Abstract

This dissertation is concerned with the enhancement of object recognition in an autonomous vehicle through advanced radar signal processing. It also explains the radar’s sequencing in importance as it is not adversely affected by weather conditions as is the case with LiDAR and cameras. Some of the developments that made this possible are the adoption of the 77 GHz frequency band, adaptive filtering and the use of machine learning. The work resolves some of the issues that include presence of interference, poor detection performance in complex environments, and high installation cost. It underscores the significance of sensing from multiple sources by describing the fusion of radar with LiDAR and cameras for a greater understanding of the environment. To this end, a series of simulations in MATLAB have been performed to support the proposed techniques and optimize the safety and reliability of AVs.