Simulations of highway chaos using fuzzy logic

We report on simulations of chaotic traffic flow in freeways based on a fuzzy knowledge based model of driver behavior. Our simulation approach is a microscopic method similar to Q. Yang and H.N. Koutsopoulos (1996), where individual vehicles are separately modeled. The knowledge base consists of an extensive and exhaustive set of fuzzy IF-THEN rules depicting all possible traffic situations that a driver is likely to encounter in a highway. The principal motivation behind using a fuzzy knowledge based approach to model the driver behavior is because fuzzy modeling provides an effective means to break up any highly nonlinear system into IF-THEN rules (J.M. Mendel, 1995). In addition, fuzzy logic is well equipped to handle uncertainties that are present in real world traffic situations (P. Chakraborty and S. Kikuchi, 1992; S. Kikuchi and M. Pursula, 1998). Therefore using fuzzy rules, it is possible to incorporate linguistic descriptions of scenarios such as `speed is moderate´ or `adjacent lane gap is quite acceptable´. Instead of quantifying variables into crisp classes as in a traditional expert system, this is the manner in which a driver is more likely to perceive any situation. Fuzzy logic techniques are receiving recent attention as important tools in transportation engineering studies and such approaches have met with a great deal of success