A fuzzy aid rear-end collision warning/avoidance system

https://doi.org/10.1016/j.eswa.2012.02.054Get rights and content

Abstract

To decrease traffic accidents is a declared target of Intelligent Transportation Systems (ITS). Among them, rear-end collisions are one of the most common and constitute one of the as yet unsolved topics in the automotive sector. This paper presents an approach to the avoidance of rear-end collisions in congested traffic situations. To this end, two fuzzy controllers, a Collision Warning System (CWS) and a Collision Avoidance System (CAS), have been developed. The former is in charge of alerting the driver in case of an impending rear-end collision to prevent or mitigate the crash. The latter is in charge of generating an output control signal for the steering wheel in order to avoid the collision. Both CWS and CAS have been tested with real cars using vehicle-to-infrastructure (V2I) communications to acquire data of vehicles. A system installed in the infrastructure capable of assessing road traffic conditions in real time is responsible for transmitting the data of the vehicles in the surrounding area. The systems have been tested at the Center for Automation and Robotics (CAR)’s facilities with two mass-produced cars.

Highlights

► We have implemented a rear-end collision warning/avoidance system in a real car. ► The system decides how to perform the maneuver without leaving the road. ► A vehicle-to-infrastructure communication system is used to exchange data. ► Fuzzy logic is used both for the warning and for the avoidance systems. ► Experiments with real cars were conducted with propper results.

Introduction

Steering distractions on the part of the driver are one of the major causes of road accidents (DGT, 2009). Since driver’s attention capacity plays a key role in the steering wheel management, a system capable of aiding the driver, would reduce significantly these accidents. Automatic steering wheel management constitutes a good option to face this problem. Presently, implementation of full automatic-steering control is one of the hardest disciplines in the intelligent-vehicles field, and perhaps has a long way to go before it comes to market (Dickmanns, 2002, Vahidi and Eskandarian, 2003). Among its possible applications, to prevent vehicle-to-vehicle collisions using steering avoidance maneuver is a very challenging topic in this field.

Vehicle-to-vehicle collisions can be classified as head-on crash, lateral impact, and rear-end collisions. The reduction in the number of two-way roads in Europe is causing rear-end collisions to occur with ever greater relative frequency. In Spain during 2008, for example, 44.3% of road accidents were rear-end collisions (DGT, 2009).

In order to reduce the number of accidents due to rear-end collisions, there have been important advances in the development of Collision Warning Systems (CWS). A model of human driving behavior focusing on the driver’s collision avoidance maneuver was presented in Kim et al. (2005). Hillenbrand, Spieker, and Kroschel (2006) developed a multilevel collision mitigation approach to obtain a solution for the decision-making problem in rear-end collisions. A longitudinal controller using terminal sliding mode with hierarchical structure in order to minimize the safety distance error and regulate the relative velocity between two vehicles was proposed for rear-end collision avoidance in Kim, Park, and Bien (2007). In a parallel line of work, vision-based systems have been developed in order to detect a potential collision situation (Chang, Tsai, & Young, 2010). Other studies have focused on the use of laser scanners, as in Kaempchen, Schiele, and Dietmayer (2009) where a collision prediction algorithm was presented based on assessing the situation as a function of Kamm’s circle. A cooperative CWS (CCWS) is presented in Sengupta et al. (2007) where five vehicles were equipped with wireless communications and global positioning systems (GPS) to know the locations and motions of all the neighboring vehicles so as to evaluate the proposed solution. Finally, Santa, Tolelo-Moreo, Zamora-Izquierdo, Übeda, and Gömez-Skarmeta (2010) presented a deep analysis about communication and navigation technology for Collision Avoidance Support Systems (CASS).

A subsequent step in rear-end collisions avoidance research was the development of automatic Collision Avoidance Systems (CAS). In this line, Adaptive Cruise Control (ACC) systems with collision avoidance capabilities using an adimensional warning index and the Time-To-Collision (TTC) were developed in Moona, Moon, and Yi (2009), where a commercial brake-by-wire system was installed in the vehicle in order to actually perform the rear-end collision avoidance. Sudeendra Kumar, Verghese, and Mahapatra (2009) implemented a CAS in an electric scaled-vehicle using a high-level protocol controller area network (CAN). The braking system of the scaled-vehicle was controlled through fuzzy logic. Vehicle collision avoidance simulation using a reactive multi-agent system in which agents interact with each other and the obstacles situated in the environment by using physics-inspired behaviors was presented in Yang, Gechter, and Koukam (2008). Finally, Labayrade, Royere, and Aubert (2007) developed a CWS based on stereovison sensors and laser scanners. When a warning message was received, a pressure command of 90 bars was applied on vehicle’s brakes.

While braking is the last response in case of an imminent collision, if the risk is detected in advance the collision should be safely avoided by a steering change. This maneuver can be incorporated both to Lane-Change maneuvers (Naranjo et al., 2008, Pérez et al., 2010a) or Lane-Departure Avoidances (Enache, Mammar, Netto, & Lusetti, 2010), but given the situation of a potential collision, the steering avoidance maneuver is more critical and more complicated than an autonomous lane-change. In this sense, an interesting study was developed in Chang and Tang (2001), which shows different strategies for drive assistance in emergency situations.

Some real applications have been implemented in collision avoidance systems. Choe, Hur, Chae, and Park (2008) use a laser range finder mounted on an experimental autonomous vehicle in an open area for the early detection of any obstacle, followed by modification of the path-plan to avoid the collision, and using a fixed safe distance of 1 m and its maximum speed is 10 km/h. Eidehall, Pohl, Gustafsson, and Ekmark (2007) developed an emergency lane assist (ELA) system to prevent dangerous lane departure maneuvers. It was mounted on a Volvo V70 equipped with forward-looking radar – for adaptive cruise control applications – and a vision system – for the lane tracking and vision-based object detection. A vehicle and steering wheel model designed to perform avoidance maneuvers was presented in Eskandarian and Soudbakhsh (2008). The effects of changing the initial velocity and the distance to the obstacle on the deviations from the desired trajectory were studied with different controllers.

Other applications, in the framework of Safety Warning Systems (SWS) have been developed with Interactive Intelligent Driver-Assistance and Safety Warning (I2DASW) and implemented in a real platform (Cheng, Zheng, Zhang, Qin, & van de Wetering, 2007) focusing on the sensor redundancy problem.

In the present study, an aid rear-end collision warning and avoidance system was evaluated and tested with real cars on a private circuit. Two fuzzy controllers were developed to implement the warning system – the trigger for the CAS and the avoidance controller. The autonomous vehicle – a convertible Citroën C3 Pluriel – receives data from the surrounding vehicles through a wireless network. The avoidance controller is responsible for taking the best action as a function of the traffic conditions, executing the avoidance maneuver in a rear-end collision risk situation.

The main contribution of this paper is the development of an aid system capable of detecting a rear-end collision situation and performing an avoidance maneuver, tested using mass-produced cars. Since CWS have been widely developed in recent years, to the best of our knowledge, the combined CWS/CAS in real roads – with real vehicles – has been not yet faced. The CAS controller is designed to perform the maneuver as soft as possible optimizing the TTC using fuzzy logic. Future trends point at wireless communications as system to detect and prevent accidents (Belanovic et al., 2010). In this sense, the data from the vehicles are acquired using a V2I system.

The structure of the paper is as follows. Section 2 describes the problem and the variables considered for the collision warning and the avoidance system. The vehicle-to-infrastructure (V2I) based communication system is presented in Section 3. The control algorithm based on fuzzy logic designed to activate the avoidance system and manage the steering wheel are described in Section 4. Section 5 describes the real cars used in the demonstrations, and the different trials at the CAR-CSIC driving circuit performed to show the feasibility of the aid system. Finally, Sections 6 Discussion, 7 Conclusions presents the discussions and conclusions drawn from the results, respectively.

Section snippets

Problem description

A rear-end collision avoidance maneuver is an emergency action involving rapid movement to negotiate an obstacle. This movement can be either a steering change or hard braking. The former can be applied when the detection is made well in advance. The latter is the last solution in case of an unavoidable accident to mitigate its consequences. The present work focuses on the former case, assuming that the detection system provides recognition of the risk situation with sufficient time.

This

V2I communication system

Wireless communication systems have been extensively used for cooperative maneuvers among cars such as ACC (Yoshinori & Yoji, 2006), Stop & Go (Naranjo et al., 2007), overtaking (Naranjo et al., 2008), intersection management (Milanés, Pérez, Onieva, & González, 2010d), and CWS (Tan & Huang, 2006), showing that it is well-suited to a wide range of real traffic situations. Due to the number of solutions presented until now, the wireless vehicle communications represents a research line by its

Control system

The control system to implement the aid maneuver is on board the vehicle. Two fuzzy controllers were developed, one for detecting a possible collision situation (CWS) and the other for performing the avoidance maneuver (CAS). The former is responsible for generating the collision warning signal so as to warn the driver via human–machine interface (HMI), and the latter generates the output control signal to manage the steering wheel in order to avoid the rear-end collision. That is, the CWS

In-circuit trials

The demonstration trials were performed at the CAR-CSIC private driving circuit using a street with a length of about 200 m.

Discussion

This paper reflects a work framed in the subject of driverless vehicles. Even if the long term goal is a utopia, we aim to obtain systems capable of aiding in driving-related tasks. In this case, we are dealing with an approach for collision avoidance systems, tested using real mass-produced vehicles. As far as we know, literature provides very little hints of real demonstrations in rear-end collision avoidance systems. Although the proposed approach provides very encouraging results, from a

Conclusions

Rear-end Collision Warning Systems have been developed by automotive manufacturers to be implemented in their commercial vehicle models in recent years, with acoustic or visual warning signals given to the driver. The next generation in safety vehicle systems is represented by Collision Avoidance Aid Systems to handle the car in order to avoid a rear-end crash.

This paper has presented a fuzzy rear-end collision avoidance system using as main inputs the time to collision and the needed

Acknowledgements

The authors are grateful to the CYCIT (Spain), Plan Nacional (Spain), and MICINN (Spain) for support under the GUIADE (P9/08), TRANSITO (TRA2008-06602-C03-01), and City-Elec (PS-370000-2009-4) projects, respectively, for the development of this work.

References (50)

  • Choe, T., Hur, J. W., Chae, J. S., & Park, Y. W. (2008). Real-time collision avoidance method for unmanned ground...
  • DGT (2009). Balance de Seguridad Vial 2008. Technical Report. Ministerio del...
  • Dickmanns, E. D. (2002). The development of machine vision for road vehicles in the last decade. In Proceedings of IEEE...
  • A. Eidehall et al.

    Toward autonomous collision avoidance by steering

    IEEE Transactions on Intelligent Transportation Systems

    (2007)
  • N.M. Enache et al.

    Driver steering assistance for lane-departure avoidance based on hybrid automata and composite lyapunov function

    IEEE Transactions on Intelligent Transportation Systems

    (2010)
  • Eskandarian, A., & Soudbakhsh, D. (2008). Enhanced active steering system for collision avoidance maneuvers. In...
  • B. Ganji et al.

    Combined quasi-static backward modeling and look-ahead fuzzy control of vehicles

    Expert Systems with Applications

    (2011)
  • Gormer, S., Muller, D., Hold, S., Meuter, M., & Kummert, A. (2009). Vehicle recognition and ttc estimation at night...
  • J. Hillenbrand et al.

    A multilevel collision mitigation approach & its situation assessment, decision making, and performance tradeoffs

    IEEE Transactions on Intelligent Transportation Systems

    (2006)
  • Horiuchi, S., Okada, K., & Nohtomi, S. (2004). Optimum steering and braking control strategies in obstacle avoiding...
  • N. Kaempchen et al.

    Situation assessment of an autonomous emergency brake for arbitrary vehicle-to-vehicle collision scenarios

    IEEE Transactions on Intelligent Transportation Systems

    (2009)
  • Kavitha, K. V. N., Bagubali, A., & Shalini, L. (2009). V2v wireless communication protocol for rear-end collision...
  • J.H. Kim et al.

    Modeling of driver’s collision avoidance maneuver based on controller switching model

    IEEE Transactions on Systems, Man, and Cybernetics, Part B

    (2005)
  • D.J. Kim et al.

    Hierarchical longitudinal controller for rear-end collision avoidance

    IEEE Transactions on Industrial Electronics

    (2007)
  • R. Labayrade et al.

    Experimental assessment of the rescue collision–mitigation system

    IEEE Transactions on Vehicular Technology

    (2007)
  • Cited by (96)

    • Traffic accidents of autonomous vehicles based on knowledge mapping: A review

      2023, Journal of Traffic and Transportation Engineering (English Edition)
    • PITS: An Intelligent Transportation System in pandemic times

      2022, Engineering Applications of Artificial Intelligence
      Citation Excerpt :

      Naranjo et al. (2003) used logic fuzzy to present an ACC (Adaptative Cruise Control) for inter-vehicle gap keeping. Fuzzy logic is also used by Montero et al. (2012) in controllers for collision avoidance or by Alzamzami and Mahgoub (2019) in Vehicular Ad hoc NETworks (VANETs). More recently, Chen et al. (2020) proposed a fuzzy Markov prediction model with the aim of estimating the short-term traffic conditions in VCPS (Vehicular Cyber–Physical Systems) in urban environment.

    • Collision-free emergency planning and control methods for CAVs considering intentions of surrounding vehicles

      2022, ISA Transactions
      Citation Excerpt :

      Li et al. [10], and Kusano et al. [11] proposed the time to collision (TTC) algorithm. Milanés et al. [12] built the time to headway (THW) model to calculate the time it takes a vehicle to reach the position of the vehicle in front of it at a certain speed. On this basis, Coelingh et al. [13] introduced the concept of relative deceleration to avoid the failure of the TTC criterion under special conditions.

    • Car crash detection using ensemble deep learning and multimodal data from dashboard cameras

      2021, Expert Systems with Applications
      Citation Excerpt :

      In particular, while the number of traffic accidents is on the relative decline, traffic accidents are still one of the leading causes of death worldwide (Heron, 2016; Sarraf & McGuire, 2020). There are two major technical solutions that help resolve this issue: collision avoidance systems (CASs) (Habib & Ridella, 2017; Milanés et al., 2012b; Yadav et al., 2020) and emergency road call services (Smirnov et al., 2013; White et al., 2011). First, a CAS, which can be installed in motor vehicles, senses the distances between vehicles through sensors, then adjusts the accelerator or automatically applies the brakes for emergency braking (Chang et al., 2010; Milanés et al., 2012a).

    View all citing articles on Scopus
    View full text