VOL. 12, NO. 14, JULY 2017 ISSN 1819-6608 ARPN Journal of Engineering and Applied Sciences ©2006-2017 Asian Research Publishing Network (ARPN). All rights reserved. www.arpnjournals.com 4270 THREAT ASSESSMENT ALGORITHM FOR ACTIVE BLIND SPOT ASSIST SYSTEM USING SHORT RANGE RADAR SENSOR Fakhrul Razi Ahmad Zakuan 1 , Hairi Zamzuri 1 , Mohd Azizi Abdul Rahman 1 , Wira Jazair Yahya 1 , Nur Hazima Faezaa Ismail 2 , Mohd Shuhari Zakariya 2 , Khairul Akmal Zulkepli 1 and Muhammad ZulfaqarAzmi 1 1 Vehicle System Engineering, Malaysia Japan International Institute of Technology, Universiti Teknologi Malaysia, Kuala Lumpur, Malaysia 2 Research Engineering Innovation Department, Perusahaan Otomobil Nasional Sdn Bhd, Shah Alam, Selangor, Malaysia E-Mail: azizi.kl@utm.my ABSTRACT Road safety has become more concern due to the number of accidents that keeps increasing every year. The safety systems include from simple installation such as seat belt, airbag, and rear camera to more complicated and intelligent systems such as braking assist, lane change assist, steering control and blind spot monitoring. This paper proposes another intelligent safety system to be implemented in passenger vehicle by monitoring the blind-spot region by using automotive short range radar as sensor to assess its surrounding. This system is called Active Blind-Spot Assist (ABSA) system and this system will collaborate with a Steering Intervention system for autonomous steering maneuvers. The objective of ABSA system is to deploy safety interventions by giving warning to the driver whenever other vehicle is detected within the blind-spot region. Furthermore, this active system also triggers autonomous steering control when the potential of collision with the detected vehicle increases greatly. Consequently, a threat assessment algorithm is developed to evaluate the right moment to give safety interventions to the driver and the conditions for autonomous steering maneuvers. The process of developing the threat assessment algorithm explained in this paper. Keywords: autonomous vehicle, blind spot assist, threat assessment, radar. INTRODUCTION Recently, road safety has become more concern as the number of accidents occur is growing. In 2014, a total of 476,196 road accidents has occurred in Malaysia alone with 6, 674 deaths [1]. Therefore, vehicle technologies are rapidly advancing worldwide specifically on safety securities [2, 3] including a passive blind-spot assist. The differences between the passive blind-spot assist system and Active Blind-Spot Assist (ABSA) system is that ABSA system is an upgraded version from the passive system. The passive blind-spot assist system is only focusing on giving warning signals to the driver. On the other hand, ABSA system is not only giving warning signals, ABSA system has authorization on maneuvering the direction of the vehicle itself. In other words, the word ‘active’ in ABSA system means that this system is capable of controlling the steering autonomously. Therefore, a threat assessment algorithm is developed for autonomous assistance to act accordingly at the right moment and situations when the system to autonomously take over control of the vehicle. The threat assessment algorithm of active blind-spot assist system or ABSA system consists of two major algorithms. The first algorithm is called collision warning algorithm, this algorithm is to warn the driver of potential collision in terms of visual and audio signals when vehicle is detected in blind-spot region. Audible and/or visible can increase the driver’s attention in such way that the driver can avoid or mitigate the collision [6]. The second algorithm is called collision avoidance algorithm for active autonomous steering controls, triggered when all the warning indicators ignored by the driver and the distance between the two vehicles get too close. A challenging aspect of the threat assessment is that it has great potential of conflicting with its main purposes. Warnings and interventions that driver consider unnecessary contribute negatively to their confidence in the ABSA system. For example, if the ABSA system gives warning signals too frequently, it will create annoyance to the driver and the driver’s trust to such system will decline greatly, and it gets worst if the steering control unexpectedly activated. Therefore, the threat assessment algorithm must be developed so that the system is able to assist the driver to ensure the vehicle safety thus increasing the system’s reliability. This system will collaborate with other autonomous systems such as lane change assist, lane keeping assist, braking assist, obstacle avoidance where ultimately will be combined to create a complete autonomous vehicle that can drive itself. Therefore, it is necessary for ABSA system to be able to provide more information as much as it can such as the target vehicle’s velocity. This paper shows the process of how the threat assessment algorithm is developed. SYSTEM HARDWARE AND DETECTION REGION ABSA system uses a couple of short range radar as sensor to monitor the blind-spot region and these radars are designed for automotive applications. They are mounted at the back corner of multi-purpose vehicle (MPV) as shown in Figure-1.