Simulação de controle adaptativo de tráfego urbano através de sistema multiagentes e com base em dados reais.

Abstract

The urban population is growing quickly in urban centers. The amount of people who use their own vehicles for transportation grows even faster. All this growth made the transport infrastructure get saturated, causing delays, congestion and increasing the time people spend in vehicles. Traffic jams have caused disorders of economic, social and environmental to the population. With the rise of the travel time and the largest number of vehicles on the streets, there are also higher levels of noise and air pollution. Moreover, in large cities, daily drivers face traffic stressful situations. Improvements in the traffic control systems have been proposed, still much remains to be done. Thus, this study evaluated the efficacy of the use of intelligent agents for the control of urban traffic in multi-agent simulation environment, in order to reduce the negative effects caused mainly by congestion. After conducting an extensive study of related work to urban traffic control and multi-agent simulation, an overview was obtained about the recent research on this subject. This overview is presented throughout the work and was the basis for the development of the proposed improvements. In this study we used the urban traffic simulator SUMO to perform the simulation of vehicular traffic in a region of Porto Alegre, Rio Grande do Sul. Actual data regarding the programming plans of existing traffic lights and about traffic volume in some city crosses were used. Then we developed an intelligent traffic light, which uses a control algorithm like a heuristic to make its decisions and the simulation was performed again. The results of two simulations with and without the proposed intelligent traffic light, were then compared and the results were considered highly relevant.