DeepAI AI Chat
Log In Sign Up

Self-organizing traffic lights at multiple-street intersections

by   Carlos Gershenson, et al.
Universidad Nacional Autonoma de Mexico

Summary: Traffic light coordination is a complex problem. In this paper, we extend previous work on an abstract model of city traffic to allow for multiple street intersections. We test a self-organizing method in our model, showing that it is close to theoretical optima and superior to a traditional method of traffic light coordination. Abstract: The elementary cellular automaton following rule 184 can mimic particles flowing in one direction at a constant speed. This automaton can therefore model highway traffic. In a recent paper, we have incorporated intersections regulated by traffic lights to this model using exclusively elementary cellular automata. In such a paper, however, we only explored a rectangular grid. We now extend our model to more complex scenarios employing an hexagonal grid. This extension shows first that our model can readily incorporate multiple-way intersections and hence simulate complex scenarios. In addition, the current extension allows us to study and evaluate the behavior of two different kinds of traffic light controller for a grid of six-way streets allowing for either two or three street intersections: a traffic light that tries to adapt to the amount of traffic (which results in self-organizing traffic lights) and a system of synchronized traffic lights with coordinated rigid periods (sometimes called the "green wave" method). We observe a tradeoff between system capacity and topological complexity. The green wave method is unable to cope with the complexity of a higher-capacity scenario, while the self-organizing method is scalable, adapting to the complexity of a scenario and exploiting its maximum capacity. Additionally, in this paper we propose a benchmark, independent of methods and models, to measure the performance of a traffic light controller comparing it against a theoretical optimum.


page 4

page 10

page 14

page 15

page 21


Modeling self-organizing traffic lights with elementary cellular automata

There have been several highway traffic models proposed based on cellula...

ELMOPP: An Application of Graph Theory and Machine Learning to Traffic Light Coordination

Traffic light management is a broad subject with various papers publishe...

Cellular Traffic Prediction and Classification: a comparative evaluation of LSTM and ARIMA

Prediction of user traffic in cellular networks has attracted profound a...

Interacting Behavior and Emerging Complexity

Can we quantify the change of complexity throughout evolutionary process...

A measure of the importance of roads based on topography and traffic intensity

Mathematical models of street traffic allowing assessment of the importa...