Fine Tuning Fuzzy Rules with Genetic programming for a fuzzy logic based Traffic light Controller System

Abstract

Fuzzy logic supports pervasive uncertainty and vagueness in real world decision making situations. Fuzzy logic traffic controller systems are investigated for their performance for number of years and given good results compared to traditional traffic controller systems. Genetic programming on the other hand is being used as a technique to automatically discover computer programs using principles of Darwinian evolution. The blend of fuzzy logic and genetic programming is used in this research to achieve fine-tuned set of fuzzy rule(s), upon giving hundreds of fuzzy rules as the input. The fuzzy logic based traffic controller system is based as the target system to retrieve simulated data, fuzzy rule sets and finally a part of the validation methodology. The system is designed based on constant average arrival rate of vehicles per lane per second, and the queue lengths of the lanes at start up of the process. Here the arrival rates are considered to be in a poisson distribution. A traffic simulator was developed with a simple interface, which handles routing of vehicles in a single four-leg intersection with left and right turns.The fuzzy logic traffic controller system is used to generate the simulation data to feeded to the genetic programming system. The genetic programming system is creating a fuzzy rule using the data feeded. We call this, fine tuning of fuzzy logic rules. This fine tuned fuzzy logic rule, proven to be resulting better Mean Square Error for queue length of the total system at a point of time.