Intelligent Traffic Light System in Bogotá for Improving Travel Times in the Transmilenio System

Abstract

This paper describes the prototype of an intelligent traffic light system that controls the green, red and yellow phases of traffic lights, with the purpose of minimizing traffic congestion in the city. The system learns traffic behavior and how to condition it based on genetic algorithms. For this purpose, the system continuously monitors traffic flow and, with this information, creates the first generations of traffic light timing. It then reviews which signal timings were best adapted to the system, that is, the ones that minimized traffic, and, finally, through crossover and mutation provided by the genetic algorithm learning method, simulating an evolutionary process, controls the city’s traffic and helps to minimize travel times in the Transmilenio mass transportation system.