Home | Issues | Profile | History | Submission | Review
Vol: 56(70) No: 2 / June 2011        

Vehicle-Following Modeling Utilizing Fuzzy Techniques
T. Oprica
University of Craiova, Automation Department, Craiova, Romania, e-mail: oprica_theodor@yahoo.com
M. Vînătoru
University of Craiova, Automation Department, Craiova, Romania, e-mail: vinatoru@automation.ucv.ro


Keywords: vehicle-following modeling, safety distance, acceleration, Sugeno fuzzy inference system, rules’ parameters, training

Abstract
It is difficult to describe driver’s behavior when following a leading vehicle, because of his individual characteristics, which give different reaction times and even decisions. Fuzzy Techniques and Artificial Neural Networks are suitable and very efficient solutions for this type of problems, which involve a certain degree of uncertainty. This paper designs a neuro-fuzzy controller, with two inputs and one output, which determines the appropriate acceleration for the following vehicle. Expected distance (DR) and relative speed (VR) are the input variables which try to express the following vehicle’s driver expectations. Subtractive clustering is applied on a training data set to create a Sugeno fuzzy inference system which describes the following vehicle’s behavior. Nine Gaussian membership functions are determined for each variable, generating nine fuzzy rules. Next, the fuzzy inference system has been introduced into a for-layers standard back-propagation neural network to tune the rules’ premise and consequent parameters by a hybrid learning algorithm. Eventually, the controller is introduced in a control system whose purpose is to assist the driver in maintaining the safety distance towards the vehicle ahead. For simulation, several different situations are taken into consider, showing the neuro-fuzzy controller’s efficiency in maintaining the safety distance towards the vehicle ahead.

References
[1] M. Pintiliei, Contribuţii privind creşterea fluenţei traficului rutier în condiţiile menţinerii şi conservării vestigiilor istorice din marile centre urbane, PhD summary, “Gh. Asachi” Technical University of Iasi, Iasi, Romania, 2009 (in Romanian).
[2] W. van Winsum, “The Human Element in Car-following Models, Transportation Research F”, vol. 2, no. 4, pp. 207-211, 1999.
[3] N. Kehtarnavaz, N. Griswold amd K. Miller, “A Transportable Neural-network Approach to Autonomous Vehicle Following,” IEEE Transactions on Vehicular Technology, vol. 47, no. 2, pp. 694-702, 1998.
[4] D. Lu, J. Rong, J. Hu and X. Liu, Modelling and Simulation of Car-Following Safety Behavior, unpublished.
[5] L. Evans and P. Wasielewski, “Risky Driving Related to Driver and Vehicle Characteristics. Accident Analysis and Prevention”, IEEE Transactions on Vehicular Technology, vol. 15, no. 3, pp. 121-136, 1983.
[6] http://www.smartmotorist.com/traffic-and-safety-guideline/maintain-a-safe-following-distance-the-3-second-rule.html.
[7] T. Oprica and M. Vînătoru, “Vehicle-following modeling utilizing neuro-fuzzy networks”, in Proc. IEEE International Joint Conference on Computational Cybernetics and Technical Informatics (ICCC-CONTI 2010), Timisoara, Romania, 2010, 6 pp.