Ghost Target Identification for Autonomotive Radar Systems
Research Poster Engineering 2025 Graduate ExhibitionPresentation by Junho Kweon
Exhibition Number 153
Abstract
Automotive radar transmits and receives signals to detect vehicles in terms of range and angle. However, reflective surfaces such as guardrails can introduce ghost targets by causing multipath propagation of radar signals. These ghost targets increase computational complexity for radar systems and may lead to incorrect detections, posing risks for autonomous vehicles. This study focuses on identifying ghost targets in automotive radar using estimated Direction-of-Departure (DOD) and Direction-of-Arrival (DOA) from received signals. To enhance detection accuracy, a physics-inspired group sparsity regularizer is introduced, and a DOD/DOA estimation algorithm is developed based on this framework. Numerical results demonstrate that the proposed algorithm outperforms existing methods in estimating DOD/DOA, which is essential in distinguishing ghost targets from real vehicles, contributing to safer and more reliable autonomous driving.
Importance
Automotive radar plays a crucial role in autonomous driving and advanced driver-assistance systems (ADAS) due to its robustness in adverse weather conditions, where cameras and LiDAR may struggle. By transmitting and receiving signals, radar detects objects based on distance and direction. However, reflective surfaces such as guardrails create multipath propagation, leading to ghost targets that increase false alarms and can result in unsafe vehicle maneuvers. Addressing this challenge is essential for improving the reliability and safety of autonomous vehicles. By effectively identifying ghost targets, this research contributes to enhancing situational awareness and decision-making for radar-equipped autonomous systems, reducing the risk of accidents and improving trust in self-driving technologies.
DEI Statement
Autonomous driving technology has the potential to improve mobility for individuals with disabilities, elderly drivers, and those with limited transportation access. My friend, who has paraplegia, faces challenges in driving, even with hand-controlled vehicles. As the aging population grows, cognitive and physical limitations increase the need for reliable autonomous solutions. Blind individuals also require transportation assistance, limiting their independence. Advancements in autonomous vehicles can provide safer and more accessible mobility solutions for these underserved populations. Automotive radar is essential for self-driving systems, but ghost targets remain a major challenge that can compromise safety. By addressing this issue, my research enhances the reliability of autonomous driving technology, benefiting those who depend on it for safe, independent transportation.