Computer simulation eases real-world Chinese traffic jams

A slight improvement in road conditions was achieved through a simulation on a general-purpose server

This photo from central Jilin City, China, shows some of the traffic congested that a computer simulation has helped ease.

This photo from central Jilin City, China, shows some of the traffic congested that a computer simulation has helped ease.

A small-scale computer simulation has reduced gridlock in central Jilin City, China, but it's no magic bullet.

NTT Data conducted the simulation of traffic conditions in Jilin using an x86 server. It was able to make recommendations that resulted in marginally better road conditions -- the average speed for vehicles in the study area improved by 1 to 2 kilometers per hour, and for public buses arrival times improved 7 percent on average.

Though modest, the results are an example of the promise of computer simulations to alleviate real-world problems in smart cities. It could have potential in massive Chinese cities such as Beijing, which is struggling with the congestion and pollution effects of over 5 million vehicles.

NTT Data had previously done a larger experiment with nine servers and more than 1 million virtual cars to simulate traffic conditions in Tokyo's 23 central wards, but it was aimed at validating the technology.

"The crucial difference is that the Jilin City experiment actually applied the result in real life, ultimately leading to alleviation of traffic congestion," a spokesman for NTT Data said via email.

In the study conducted between November and December, NTT Data gathered GPS data from terminals installed on about 200 buses circulating on eight routes. It also looked at road layouts, regulations such as no left turns, traffic volume information and traffic light data from about 10 intersections.

The data were fed into the simulation, which is a virtual road network with multiple "agents" representing vehicles, traffic signals, roads and intersections. The simulation also features algorithms developed by NTT Data to forecast traffic congestion.

A server equipped with a 12-core Intel Xeon E5-2697 v2 processor ran S-quattro, a general-purpose simulation program developed by NTT Data Mathematical System. The calculation was performed in under an hour.

The simulation examined parameters including how long traffic lights at a given intersection go through one green-yellow-red cycle as well as the time between a green light at one signal and a green light at the next. It tried various combinations of traffic light settings to find the result that eased the flow of traffic without making congestion worse.

The best recommendations were implemented at six intersections in central Jilin City, a rapidly growing metropolis of over 4 million in China's northeast, resulting in the improvements. The city plans to add more intersections using the simulation's recommendations, and gather more traffic data based on GPS-equipped taxis.

Tim Hornyak covers Japan and emerging technologies for The IDG News Service. Follow Tim on Twitter at @robotopia.

Join the newsletter!

Or

Sign up to gain exclusive access to email subscriptions, event invitations, competitions, giveaways, and much more.

Membership is free, and your security and privacy remain protected. View our privacy policy before signing up.

Error: Please check your email address.

Tags internetNTT DataInternet-based applications and services

More about IDGIntelNews

Show Comments
[]