Chaos Control of Freeway Mainline Using Variable Speed Limits with Fuzzy-Neural Networks Based on Subtractive Clustering

Abstract

The chaos control of freeway mainline was studied by using variable speed limits and fuzzy-neural networks (FNNs) based on subtractive clustering. Based on the uncertainty and nonlinearity of a traffic system, the establishment of a knowledge base of a mainline chaos controller for freeway was proposed by using data mining technology. The chaos control principle of mainline variable speed limits in freeway was briefly introduced. The Takagi-Sugeno FNNs chaos controller was designed, where traffic density, upstream traffic volume, and maximal Lyapunov exponent were selected as the input variables, whereas mainline speed upper limit was selected as the output variable of the controller. Subtractive clustering was used to determine the controller structure, including the extraction of fuzzy rules and generation of initial parameters. The radius of the clustering centers was optimized using the genetic algorithm, and the parameters of the fuzzy controller were optimized using FNN. The simulation result indicated that order motion on freeway can be realized by using the mainline intelligent chaos controller designed based on the proposed method to suppress traffic jam and to enhance traffic volume.