Uncorrected Author Proof Journal of Intelligent & Fuzzy Systems 39 (2020) 1–18 DOI:10.3233/JIFS-190634 IOS Press 1 A lightweight vehicle detection and tracking technique for advanced driving assistance systems 1 2 3 Wael Farag a,b,∗ 4 a College of Engineering and Technology, American University of the Middle East, Kuwait 5 b Department Electrical Engineering, Cairo University, Egypt 6 Abstract. In this paper, an advanced-and-reliable vehicle detection-and-tracking technique is proposed and implemented. The Real-Time Vehicle Detection-and-Tracking (RT VDT) technique is well suited for Advanced Driving Assistance Systems (ADAS) applications or Self-Driving Cars (SDC). The RT VDT is mainly a pipeline of reliable computer vision and machine learning algorithms that augment each other and take in raw RGB images to produce the required boundary boxes of the vehicles that appear in the front driving space of the car. The main contribution of this paper is the careful fusion of the employed algorithms where some of them work in parallel to strengthen each other in order to produce a precise and sophisticated real-time output. In addition, the RT VDT provides fast enough computation to be embedded in CPUs that are currently employed by ADAS systems. The particulars of the employed algorithms together with their implementation are described in detail. Additionally, these algorithms and their various integration combinations are tested and their performance is evaluated using actual road images, and videos captured by the front-mounted camera of the car as well as on the KITTI benchmark with 87% average precision achieved. The evaluation of the RT VDT shows that it reliably detects and tracks vehicle boundaries under various conditions. 7 8 9 10 11 12 13 14 15 16 17 18 Keywords: Computer vision, self-driving car, autonomous driving, ADAS, vehicle detection, vehicle tracking 19 1. Introduction 20 Increasing safety, reducing road accidents and 21 enhancing comfort and driving experience are the 22 major motivations behind equipping modern cars 23 with Advanced Driving Assistance Systems (ADAS) 24 [1]. In the past couple of decades, major car manufac- 25 turers introduce many sophisticated ADAS functions 26 [3] like Electronic Stability Control (ESC), Anti- 27 lock Brake System (ABS), Lane Departure Warning 28 (LDW) [5], Lane Keep Assist (LKA) [6], etc. 29 ∗ Corresponding author. Wael Farag, E-mail: wael.farag@aum.edu. kw, wael.farag@cu.edu.eg. These functions represent steady incremental steps 30 toward a hypothetical future of safe fully autonomous 31 vehicles [7]. 32 Most recent ADAS functions like Collision Avoid- 33 ance, Automated Highway Driving (Autopilot), 34 Automated Urban Driving, Automated Parking and 35 Cooperative Maneuvering require more and more fast 36 and reliable detection and tracking for on-road vehi- 37 cles [12], which is among the most complex and 38 challenging tasks. In order to successfully detect the 39 other vehicles on the road, accurate localization of 40 potential vehicles in camera images or LiDAR data 41 is required, the relative position of these cars with 42 respect to the road needs to be determined, and the 43 ISSN 1064-1246/20/$35.00 © 2020 – IOS Press and the authors. All rights reserved