, pp. 1-10 http://dx.doi.org/10.14257/ijseia.2015.9.7.01 Study on V2V-based AEB System Performance Analysis in Various Road Conditions at an Intersection Sangduck Jeon 1, Gyoungeun Kim 1 and Byeongwoo Kim 2 1 Graduate School of Electrical Engineering, University of Ulsan, 93 Daehak-ro, Ulsan, Republic of Korea jsd0831@gmail.com, gyg509@gmail.com 2 School of Electrical Engineering, University of Ulsan, 93 Daehak-ro, Ulsan, Republic of Korea bywokim@ulsan.ac.kr Abstract This paper proposes an autonomous emergency braking (AEB) system that takes into account road conditions at an intersection, using vehicle-to-vehicle (V2V) communication. The conventional AEB system considers the friction characteristics of a particular type of road, and thus has a fixed timing for activating the automatic braking without regard to different road conditions. Such an AEB intervention feature has performance limitations in actual road conditions. To overcome these drawbacks, this study proposes a control method that adjusts the automatic brake application time to road conditions. Additionally, collision risks at an intersection were calculated using various road friction coefficients and the V2V-based speed inputs from adjacent vehicles. The efficacy of the proposed AEB system was validated through tests under various scenarios, applying the road friction coefficient and vehicle speed as variables. The tests verified the improved collision avoidance performance of the proposed AEB system s control method compared to the conventional method. Keywords: Road condition, Autonomous Emergency Braking (AEB), Vehicle to Vehicle (V2V), Vehicle to Infra (V2I), Time to Collision (TTC), Collision Avoidance 1. Introduction Vehicle electronics, and the digitalization of automotive technologies, is a fastgrowing segment of the automotive market. Every year, new electronic devices are presented for enhanced driver convenience and safety. Extensive studies are underway worldwide in this area, revolving around the Advanced Driver Assistance System (ADAS), a generic term for designating cutting-edge driver-support systems to enhance driver and road safety [1, 2]. An autonomous emergency braking (AEB) system, which analyzes forward collision risks and applies automatic braking to prevent crashes, is central to the ADAS [3]. Currently available AEB systems detect obstacles using radar and camera sensors [4]. Physically obstructed by adjacent vehicles or other obstacles, radar and camera sensors can detect only the objects that are directly in front of them. In particular, due to this limitation of obstacle perception sensors, crash risks are higher at an intersection where vehicles pass in various directions and the view is obstructed by blind spots [5, 6]. To overcome this problem, vehicle-to-infrastructure (V2I) and vehicle-to vehicle (V2V) communication technologies were introduced into the AEB system [7-10]. The performance of AEB systems is generally evaluated using the Euro NCAP AEB evaluation procedure as the currently valid objective rating procedure. In ISSN: 1738-9984 IJSEIA Copyright c 2015 SERSC
accordance with this procedure, previous studies on AEB control methods have uniformly considered only dry asphalt [11]. Road features as perceived by the driver and the road friction characteristics among the information input while driving are essential in the driving process and can be useful for vehicle control systems. Therefore, an AEB system reflecting only particular road conditions cannot be expected to satisfy efficient AEB system intervention in constantly changing road conditions [12]. Against this background, this study presents a new AEB system that reflects road friction conditions at an intersection on the basis of V2V communication. Its efficacy is verified by comparing its collision avoidance performance with that of a conventional system that does not consider the road friction conditions as a variable. 2. System Design Figure 1 presents the AEB system block proposed in this paper that considers road conditions. It consists of two functional domains: 1) calculation of the time to collision (TTC), which represents the collision risk at an intersection, and 2) calculation of the time for emergency brake application, considering the road friction coefficient. Figure 1. Proposed AEB System Configuration The following process shows TTC calculation under intersection conditions. The proposed AEB system activation time is the moment in which the collision risk of the host vehicle ( TTC Host ) becomes smaller than the brake application time, according to the road surface conditions (TTC propose ), as expressed by Eq. (1). Full braking is applied once the proposed AEB system is activated. TTC Host < TTC propose (Full braking) (1) We calculated the collision risk at an intersection using V2V communication [13]. In order to estimate the collision risk of the host vehicle ( TTC Host ) with the target vehicle at the expected collision point of the intersection, relative distance (RD) and relative angle (RA) were the considered variables, as shown in Figure 2. RD and RA denote the straight line between the host and target vehicles and the angle this line forms with the line from the host vehicle to the expected collision point, respectively. They are obtained through the V2V information input from the adjacent vehicles and the host vehicle. 2 Copyright c 2015 SERSC
The distances from the host vehicle and the target vehicle were calculated using Eqs. (2) and (3), respectively, using RD and RA. TTC Host and TTC Target were then calculated using Eqs. (4) and (5), respectively. If the absolute difference obtained by subtracting TTC Target from TTC Host was smaller than the threshold, it was assumed that the two vehicles would collide at the intersection. The level of risk was determined by tracking the changes in the TTC Host. Figure 2. Calculation of the Expected Collision Point using V2V Communication Intersection distance Host = RD cos (RA) (2) Intersection distance Target = RD sin (RA) (3) TTC Host = Intersection distance Host Vehicle speed Host (4) TTC Target = Intersection distance Target Vehicle speed Target (5) TTC Host TTC Target < threshold (6) An AEB system estimates the forward collision risk and automatically activates the braking system if a collision is judged imminent or unavoidable. If the road is wet or icy, the friction between the tires and the road is smaller so that system intervention should occur faster. The brake application timing that considers road conditions, as proposed in this paper, is calculated as follows. Eq. (7) is a formula for calculating the brake application time (TTC propose ), according to road conditions, using the speed and deceleration time to stop. Brake distance (S μ ) and time (T μ ) from brake application to stop can be calculated from Eqs. (8) and (9), respectively, using maximum deceleration. Eq. (10) is the formula for the maximum possible deceleration (a μ ). Copyright c 2015 SERSC 3
TTC propose = S μ V Host (7) S μ = V Host T μ 2 (8) T μ = V Host a μ (9) a μ = μ g (10) 3. Simulation and Results 3.1. Simulation Scenario Figure 3 presents the simulation environment defined to validate the AEB system proposed in this paper. The driver of the host vehicle cannot see the vehicle approaching the intersection from the other direction because of the blind spot created by buildings, which leads to the collision between the host and target vehicles at the intersection. Figure 3. Simulation Configuration Table 1 provides an overview of the simulation scenarios. In compliance with the Euro NCAP AEB test scenario, 40[km/h] (representing the speed in a city) and 60[km/h] (representing the inter-urban driving speed) were set as representative vehicle speeds. Three different road conditions were configured according to their respective friction coefficients: dry asphalt (μ = 0.85), wet asphalt (μ = 0.60), and snow asphalt (μ = 0.30). Although the test setting in the Euro NCAP AEB evaluation procedure involves only dry asphalt for conventional AEB systems, we tested the AEB system under all three road conditions to compare the collision avoidance performance of the conventional and proposed AEB systems. Simulations were performed at two vehicle speeds under three road conditions in accordance with the test design. 4 Copyright c 2015 SERSC
Table 1. Simulation Scenarios AEB scenario Speed [km/h] Road surface condition Friction factor City 40 Dry asphalt μ = 0.85 Inter-Urban 60 Wet asphalt μ = 0.60 Snow asphalt μ = 0.30 3.1. Simulation Results Table 2 shows the emergency brake application timing for different road conditions when decelerated from the traveling speed of 60[km/h] for both the host and target vehicles. Although the conventional AEB system started the braking intervention and applied full braking at the same time points (1.6[s] and 0.7[s], respectively) for all three road conditions, the intervention onset of the proposed AEB system varied according to road conditions (1.00, 1.42, and 2.83[s] for dry, wet, and snow asphalt). Table 2. Emergency Braking Intervention Timing of the Conventional and Proposed AEB Systems Under Different Road Conditions Dry asphalt (μ = 0.85) Wet asphalt (μ = 0.60) Snow asphalt (μ = 0.30) TTC[s] TTC[s] TTC[s] Conventional 1.60 1.60 1.60 0.70 0.70 0.70 Proposed 1.00 1.42 2.83 Table 3 shows the results (collision or avoidance) of the simulations performed according to the test settings presented in Table 1. As Table 3 illustrates, the host vehicle with the conventional AEB system avoided collision with the target vehicle only on the dry asphalt. In contrast, the proposed AEB system avoided collision under all three road conditions, thus proving the proposed AEB system as an efficient brake application timing calculation algorithm. Table 3. Collision/avoidance according to Speed and Friction Factor AEB system Conventional Proposed Road condition Speed [km/h] Dry asphalt (μ = 0.85) Wet asphalt (μ = 0.60) Snow asphalt (μ = 0.30) 40 Avoided Collided Collided 60 Avoided Collided Collided 40 Avoided Avoided Avoided 60 Avoided Avoided Avoided Figure 4 and Figure 5 illustrate the acceleration of the vehicle with the conventional AEB system at the traveling speed of 60[km/h]. In order to verify the impact of road conditions on braking performance, a braking test was performed with dry and snow asphalt as typical contrast conditions. The acceleration on the dry asphalt was 0.4 g for partial braking and 0.85 g for full braking. In contrast, the acceleration on the snow asphalt was 0.3 g for both partial and full brake applications. It was thus verified that, unlike on the dry asphalt, partial and full brake applications cannot be implemented on snow asphalt. This information proves that road conditions influence braking performance substantially, implying that AEB Copyright c 2015 SERSC 5
systems should take into account the road friction factor because earlier emergency brake application is required as the road friction force decreases. Figure 4. Acceleration at Dry Asphalt and 60[km/h] Figure 5. Acceleration at Snow Asphalt and 60[km/h] We selected a vehicle speed of 60[km/h] traveling on the wet asphalt as a typical setting in which the efficacy of the proposed AEB system can be demonstrated. Figure 6 and Figure 7 illustrate the TTC and vehicle velocity changes under the selected conditions. The performance of the proposed AEB system can be analyzed and evaluated with the test results. Figure 6 shows the analysis results of the changes in the TTC, which represents the collision risk relative to time, thus comparing the braking performance of the conventional and proposed systems. As Figure 6 shows, the conventional AEB system applied braking at 1.6[s], which resulted in collision at 6.4[s]. In contrast, the proposed AEB system made efficient use of the distance remaining to a full stop and avoided collision by applying full breaking at 1.42[s], as calculated by the 6 Copyright c 2015 SERSC
proposed AEB algorithm reflecting the road friction coefficient under wet conditions. Figure 6. TTC Changes at Wet Asphalt and 60[km/h] Figure 7 shows the velocity changes of the host vehicle on the time axis. The conventional AEB system applied partial braking between 4.2[s] and 5.5[s] and full braking afterwards, but could not avoid collision, with the velocity at the moment of collision being 20[km/h]. In contrast, the proposed AEB system applied full braking at 4.5[s] and avoided collision. The conventional AEB system failed because it logically does not reflect the road friction coefficient. Figure 7. Velocity Changes at Wet Asphalt and 60[km/h] Copyright c 2015 SERSC 7
4. Conclusions This paper presented an improved strategy for the braking control of the AEB system by adjusting the emergency brake application timing to road conditions using V2V communication. The proposed AEB system estimated collision risk at an intersection using the V2V-based information input regarding locations and velocities of the ego and adjacent vehicles. The brake application timing was calculated using the time required for the deceleration to come to a full stop according to the road friction conditions and the velocity of the ego vehicle. Emergency brake application time is finally determined on the basis of the estimated collision risk and the required braking time according to road conditions. Simulation testing verified that the braking performance of the conventional AEB system was affected by road conditions at a travel speed of 60[km/h]. On the dry asphalt, partial and full brake application occurred with accelerations of 0.4 g and 0.85 g, respectively, at TTCs of 1.6[s] and 0.7[s], thus avoiding collision. However, on the snow asphalt, braking with acceleration beyond 0.3 g could not be implemented because of the decreased road friction force, resulting in collision. The emergency brake application timing of the conventional AEB system is fixed at the TTCs of 1.6[s] and 0.7[s] for partial and full brakes, respectively, because its calculation is based on the driving conditions on dry asphalt without considering other road conditions. As a result, the testing resulted in collision on wet and snow asphalt. In contrast, the proposed AEB system succeeded in avoiding collision at an intersection in all test conditions of velocity and road conditions, thus demonstrating its improved braking performance compared to the conventional one. In follow-up studies, the performance of collision detection systems will be tested on various intersection types other than the right-angle intersection. Acknowledgements This research was supported by the Ministry of Science, ICT and Future Planning (MSIP), Korea, under the Convergence Information Technology Research Center (C- ITRC) (IITP-2015-H8601-15-1005), and supervised by the Institute for Information & Communications Technology Promotion (IITP). References [1] D. Park and Y. Kwak, Analysis and Design of E-Band Microstrip Array Antennas for an Adaptive Cruise Control System, IJCA, (2013). [2] X. Yin and M. Wang, Research on Safety Distance Mathematical Model of Pro-Active Head Restraint in Rear-End Collision Avoidance System, IJSIA, (2015). [3] Automated Emergency Brake Systems: Technical Requirements, Costs and Benefits, Published Project Report, European Commission, (2008). [4] Comparative Test of Advanced Emergency Braking Systems, Test report, ADAC Vehicle Testing, (2013). [5] Road Traffic Authority Driver s License Examination Office, http://www.koroad.or.kr/. [6] Institute of Transportation Engineers: Intersection Safety Briefing Sheet, U.S. Department of Transportation Federal Highway Administration, (2004). [7] F. Basma, Y. Tachwali and H. H. Refai, Intersection Collision Avoidance System Using Infrastructure Communication, 14th International IEEE Conference on Intelligent Transportation Systems, (2011). [8] A. Eichhorn, P. Zahn and D. Schramm, A Warning Algorithm for Intersection Collision Avoidance, Springer International Publishing, (2013). [9] J. Ibanez-Guzman, S. Lefevre, A. Mokkadem and S. Rodhaim, Vehicle-to-Vehicle Communications Applied to Road Intersection Safety, Field Results, 2010 13th International IEEE Conference on Intelligent Transportation Systems, (2010). [10] S. Jeon, G. Kim and B. Kim, Braking Performance Improvement Method for V2V Communication- Based Autonomous Emergency Braking at Intersections, Advanced Science and Technology Letters, Ubiquitous Science and Engineering, (2015). [11] Euro NCAP AEB Test Protocol. EURO NCAP. http://www.eruoncap.com, (2013). 8 Copyright c 2015 SERSC
[12] R. Heseinnezhad and A. Bab-Hadiashar, Efficient Antilock Braking by Direct Maximization of Tire- Road Frictions, IEEE Transaction on Industrial Electronics, (2010). [13] H. Cho and B. Kim, Study on Cooperative Intersection Collision Detection System Based on Vehicleto-Vehicle Communication, Advanced Science and Technology Letters, Electrical Engineering, (2014). Authors Sangduck Jeon, received the B.S. Degree in elecrical engineering from Ulsan University, Ulsan, South Korea, in 2014. Where he is currently working toward the M.S. degree in automotive Electronics and control laboratory from Ulsan University, Ulsan, South Korea. His main research activities and interests are autonomous car and autonomous emergency braking (AEB) system. Gyoungeun Kim, received the B.E degree in electrical engineering in University of Ulsan, Ulsan, South Korea. She is pursuing her integrated master and Ph.D degree in electrical engineering in University of Ulsan, Ulsan, South Korea. Her current research interests include advanced driving assistance system (ADAS), path planning and autonomous emergency braking (AEB) system. Byeongwoo Kim, received the B.E, M.E. and Ph.D degree in Precision Mechanical Engineering from Hanyang University. He worked at KOSAKA Research Center in 1989. He worked at KATECH electrical technology Research Center from 1994 to 2006. Now he is a professor in the School of electrical engineering in University if Ulsan, Ulsan, South Korea from 2006. His current research interests include advanced driving assistance system (ADAS), and autonomous emergency braking (AEB) system. Copyright c 2015 SERSC 9
10 Copyright c 2015 SERSC