This research work presents the previous results of implementing autonomous traffic light control system based on sophisticated agents to overcome problems like congestion, pollutant emissions and fuel consumption in modern cities. The proposed agent based approach uses back propagation neural networks to provide green light intervals according to the demand level of the intersection. The effectiveness of this proposal is tested simulating two traffic intersections.
To do this, the paper also introduces a novel simulator and emission analyzer developed to run a set of tests in order to compare the presented methodology with traditional traffic control methods. Preliminary results demonstrate the efficiency of the introduced approach, offering significant mobility and environmental benefits. For example, for the first test and using observed traffic volumes, our approach increase mobility in 28% and reduce the fuel consumption in 20%.