Research of optimizing control based on rough sets for urban traffic signal

In order to gain better dynamic response and enhance the real-time character of control system, a novel neural fuzzy approach based on rough sets is proposed to control traffic signals at an intersection in this paper. The approach generates experiential rules from real sample data with rough sets theory assisted feature reduction method and performs a rough sets based autotuning for the neural fuzzy controller. The rule generation mechanism maintains the underlying semantics of the feature set. Research findings in the study indicate that the use of rough set theory to aid the neural fuzzy controller can produce relatively better control performances and the approach is effective and practical for traffic signals control.