Density-Based Vehicle Detection Approach for Automotive Millimeter-Wave Radar

Automotive radars, along with other sensors, generate the backbone of self-driving vehicles. Herein, automotive radars, especially the millimeter-wave (MMW) radar, have already reached a market penetration that leads to tens of million units being used. The MMW radar has been rapidly expanded and developed for the past a few years, and it has found its own way into nearly all car manufacturers' plans in the world. In this paper, we focus on the classic signal processing problem, vehicle detection, based on the MMW radar. It is a new area for MMW radar signal processing and requires a both effective and efficient solution for practical applications, like the advanced driver assistant system (ADAS). We first generate high-resolution point cloud image based on radar signal processing steps and develop two kinds of density-based approaches for vehicle detection tasks on this point cloud image. The article outlines the processing steps and presents some experimental results.