Self-organizing traffic lights by Carlos Gershenson
Pitch
Traffic cannot be predicted. Let traffic lights adapt themselves by reacting to their current local demand.
Description
Guidance on collaborative pilot
This is a pilot test of a new, collaborative approach for getting work done in the Climate CoLab. It will run during March and April of 2012.Just like in the 2011 activities, anyone can create a proposal. But there is also a community proposal, where members are encouraged to work together in a collaborative way. Any member can contribute to the community proposal as long as they are logged in.
The community proposal is like a wiki, so the history of edits is tracked, and you can revert to prior versions of the proposal if desired.
Please also use the Comments to express your opinion on whether or not you would like to see this collaborative approach used in the Climate CoLab in 2012.
Feel free to organize the proposal as you see fit. One thought—it's good to have a brief summary of the overall proposal at the top, as an aid to readers.Proposal Text
Summary
The optimal coordination of traffic lights is an extremely complex problem. Moreover, traffic situations change constantly, demanding everchanging solutions. Most traffic lights are fixed. And from the few ones that adapt, they do so very slowly. A recently proposed method allows for the distributed adaptaiton of traffic lights as fast as the traffic demands change, i.e. at the seconds scale.
In several computer simulations, this method has proven to reduce waiting times by 50%, leading to considerable emission reductions. For example, it is estimated that with one thousand intersections in Mexico City, one million CO2 tons would be saved every year, at a cost of only $25 million.
We have the technology to implement this solution, it is time to do it!
Category of the action
Reducing emissions from transportation
What actions do you propose?
Installing sensors and controllers at traffic light intersections in cities to improve traffic flow.
This method can seamlessly give priority to public transport, emergency vehicles, etc, without disrupting the rest of the network.
Who will take these actions?
Local goverments, city councils.
Where will these actions be taken?
We have the agreement to make a pilot study within the main campus of the Universidad Nacional Autonoma de Mexico (UNAM). We recently contacted people from the Mexico City government and they are also interested in a pilot study. Alternative locations can be considered in Mexico and in other countries.
How much will emissions be reduced or sequestered vs. business as usual levels?
In Cools et al. (2007), in a simulation of an avenue in Brussels with an early version of the method, we estimated yearly reductions of more than 1300 tons of CO2, 4400kg of NOx, 66000kg of CO, and 9000kg of CxHx implementing only ten intersections within 1km.
Extrapolating these results, we estimate a yearly reduction of one million tons of CO2 for the main 1000 intersections in Mexico City. Similar savings can be estimated depending on the city size and average travel distances.
What are other key benefits?
In several studies we have found average reductions of waiting times of 50%, an estimated reduction of total travel times of 25%. This would not only save millions of person-hours per day, but reduce traffic jams and health issues related to traffic, improving tangibly the quality of life of citizens.
What are the proposal’s costs?
Currently, the equipment required to implement self-organizing traffic lights is about $25,000 per intersection (e.g. ITERIS Vantage® RZ-4 Camera, Edge 2 processor, and VRack, with a SBC-2400 Traffic Controller running Wapiti controller software). We would like to implement a pilot study in the main campus of UNAM, where there are eight traffic lights, including pedestrians and bus rapid transit. Considering $100,000 for hiring experts to work full time on this project, the cost of this pilot study would be $300,000.
This cost considers only sensors and controllers. Intersections already have modern traffic lights, and the method does not require any communication between intersections, as all decisions are taken with local information.
With mass production, the cost of the sensors and controllers could be reduced considerably, to the order of $2000-$4000 per intersection.
There is also the intention of a spin-off company to produce and install these sensors and controllers worldwide at a lower cost. Thus, this project can be interesting to venture capitalists as well.
Time line
The pilot study could be implemented and analyzed in twelve months, offering complementary results to those already obtained with computer simulations.
The pilot study would show more clearly and in a real situation the benefits of self-organizing traffic lights, propagating their implementation worldwide.
Related proposals
Lämmer and Helbing (2008) proposed a similar method for coordinating traffic lights, where independent traffic lights use sensors to adapt to current demands. However, there are key differences, e.g. their proposal adjusts cycle times, while in the present one all light changes are on demand. These differences are considerable enough to implement this project without infringing their patent, and also because of the time in which our work has been published.
We have not compared our methods directly, but from their reported results, ours offers greater improvements over existing configurations.
References
- Ball, P. (2004). Beating the lights. News@Nature. http://dx.doi.org/10.1038/news041129-12
- Cools, S. B., Gershenson, C., and D’Hooghe, B. (2007). Self-organizing traffic lights: A realistic simulation. In Prokopenko, M., editor, Self-Organization: Applied Multi-Agent Systems, chapter 3, pages 41–49. Springer. http://arxiv.org/abs/nlin/0610040
- Gershenson, C. (2005). Self-organizing traffic lights. Complex Systems, 16(1):29–53. http://www.complex-systems.com/pdf/16-1-2.pdf
- Gershenson, C. (2007). Design and Control of Self-organizing Systems. CopIt Arxives, Mexico. http://tinyurl.com/DCSOS2007
- Gershenson, C. (2012). Self-organizing urban transportation systems. In Portugali, J., Meyer, H., Stolk, E., and Tan, E., editors, Complexity Theories of Cities Have Come of Age: An Overview with Implications to Urban Planning and Design, pages 269–279. Springer, Berlin Heidelberg. http://arxiv.org/abs/0912.1588
- Gershenson, C. (2013). Living in living cities. Artificial Life, In Press. http://arxiv.org/abs/1111.3659
- Gershenson, C. and Rosenblueth, D. A. (2012a). Adaptive self-organization vs. static optimization: A qualitative comparison in traffic light coordination. Kybernetes, 41(3):386–403. http://www.emeraldinsight.com/journals.htm?articleid=17035790
- Gershenson, C. and Rosenblueth, D. A. (2012b). Self-organizing traffic lights at multiple-street intersections. Complexity, 17(4):23–39. http://arxiv.org/abs/1104.2829
- Lämmer, S. and Helbing, D. (2008). Self-control of traffic lights and vehicle flows in urban road networks. J. Stat. Mech., P04019. http://arxiv.org/abs/0802.0403
By:
Carlos Gershenson
