Vehicle Detection for Autonomous Driving: A Review of Algorithms and Datasets

• Published in: IEEE transactions on intelligent transportation systems Vol. 24; no. 11; pp. 1 - 27
• Main Authors: Karangwa, Jules, Liu, Jun, Zeng, Zixuan
• Format: Journal Article
• Language: English
• Published: New York IEEE 01.11.2023; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Algorithms; Automobiles; Autonomous driving; Autonomous vehicles; camera; Datasets; Driving; Feature extraction; LiDAR; Machine vision; Millimeter waves; mmWave radar; Object detection; sensor fusion; Task analysis; Three-dimensional displays; Vehicle detection; Vehicles
• ISSN: 1524-9050; 1558-0016
• DOI: 10.1109/TITS.2023.3292278

Nowadays, vehicles with a high level of automation are being driven everywhere. With the apparent success of autonomous driving technology, we keep working to achieve fully autonomous vehicles on roads. Efficient and accurate vehicle detection is one of the essential tasks in the environment perception of an autonomous vehicle. Therefore, numerous algorithms for vehicle detection have been developed. However, their strengths in terms of performance have not been deeply assessed or highlighted yet. This work comprehensively reviews the existing methods and datasets for vehicle detection considering their performances and applications in the field of autonomous driving. First, we briefly describe tasks, evaluation criteria, and existing public datasets for vehicle detection in autonomous driving. Second, we provide a rigorous review of both classical and latest vehicle detection methods, including machine vision-based, mmWave radar-based, LiDAR-based, and sensor fusion-based methods. Finally, we analyze the pertinent challenges of autonomous vehicles and provide recommendations for future works concerning vehicle detection. The present review covers over 300 research works and aims to help researchers interested in autonomous driving, especially in vehicle detection.