Fuzzy Control For Unban Freeway Ramp Metering And Controller Optimization Based On Genetic Algorithm

This dissertation begins with the analysis of freeway problems, and freewaycongestion is the worst problem of freeway. The best solution for this problem iscontrol and manage freeway. In all control strategies, ramp metering is one of themost realistic and sophisticated ways and wildly used. This dissertation classifies andsummarizes these existing ramp metering algorithms. It analyses these mathematicmodels based on control techniques and the complex of freeway. The freeway's complexity, nonlinear nature and non-stationary behavior makesan accurate system model is unavailable so a suitable fuzzy controller having adistinct advantage over traditional controllers applied. The inputs of this controller areupstream flow and ramp section density, the output is ramp metering rate. Fuzzy ramp control acquires good result, but fuzzy rules coming from experienceis its shortcoming. The selection of fuzzy membership functions' parameters willaffect the result. To overcome the conventional problems of the calibration process offuzzy controllers and improve the overall performance of ramp metering, an geneticalgorithm(GA) is integrated into the system to tune parameters of fuzzy controller. The results of simulation illustrate the superiority of the fuzzy control to theoptimal control for the changeful traffic condition. In the case of the whole trafficcondition change in a large range or the traffic in some road becoming heavy, the GAcan adapt the parameters of all the ramp fuzzy controllers, keep the traffic running in adesired condition.