Using Machine Learning to Detect Ghost Images in Automotive Radar

Abstract

Radar sensors are an important part of driver assistance systems and intelligent vehicles due to their robustness against all kinds of adverse conditions, e.g., fog, snow, rain, or even direct sunlight. This robustness is achieved by a substantially larger wavelength compared to light-based sensors such as cameras or lidars. As a side effect, many surfaces act like mirrors at this wavelength, resulting in unwanted ghost detections. In this article, we present a novel approach to detect these ghost objects by applying data-driven machine learning algorithms. For this purpose, we use a large-scale automotive data set with annotated ghost objects. We show that we can use a state-of-the-art automotive radar classifier in order to detect ghost objects alongside real objects. Furthermore, we are able to reduce the amount of false positive detections caused by ghost images in some settings.

References in corpus (2)

• Seeing Around Street Corners: Non-Line-of-Sight Detection and Tracking In-the-Wild Using Doppler Radar
• Automated Ground Truth Estimation For Automotive Radar Tracking Applications With Portable GNSS And IMU Devices

Cited by in corpus (5)

• Anomaly Detection in Autonomous Driving: A Survey
• The Radar Ghost Dataset -- An Evaluation of Ghost Objects in Automotive Radar Data
• BoostRad: Enhancing Object Detection by Boosting Radar Reflections
• Object Modeling from Underwater Forward-Scan Sonar Imagery with Sea-Surface Multipath
• Reflection-Aware Reasoning for Non-Line-of-Sight Pedestrian Localization