Search the Community
Showing results for tags 'signalguru'.
Found 1 result
Traffic management APPlied logic Sep 13th 2011, 16:10 by The Economist online TRAFFIC lights are crucial tools for regulating traffic flow. They are not, however, perfect. Drivers exchange the gridlock that would happen at unmanaged junctions for a pattern of stop-go movement that can still be frustrating, and which burns more fuel than a smooth passage would. Creating such a smooth passage means adjusting a vehicle’s speed so that it always arrives at the lights when they are green. That is theoretically possible, but practically hard. Roadside signs wired to traffic lights can help get the message across a couple hundred metres from a junction, but such signs are expensive, and have not been widely deployed. Margaret Martonosi and Emmanouil Koukoumidis at Princeton University, and Li-Shiuan Peh at the Massachussets Institute of Technology, however, have an idea that could make the process cheaper and more effective. Instead of a hardwired network of signs, they propose to use mobile-phone apps. For a driver to benefit, he must load the team’s software, dubbed SignalGuru, into his phone and then mount it on a special bracket attached to the inside of his car’s windscreen, with the camera lens pointing forwards. SignalGuru is designed to detect traffic lights and track their status as red, amber or green. It broadcasts this information to other phones in the area that are fitted with the same software, and—if there are enough of them—the phones thus each know the status of most of the lights around town. Using this information, SignalGuru is able to calculate the traffic-light schedule for the region and suggest the speed at which a driver should travel in order to avoid running into red lights. Tests in Cambridge, Massachusetts, where five drivers were asked to follow the same route for three hours, and in Singapore, where eight drivers were asked to follow one of two routes for 30 minutes, revealed that SignalGuru was capable of predicting traffic-light activity with an accuracy of 98.2% and 96.3% respectively, in the two cities. This was particularly impressive because in Cambridge the lights shifted, roughly half-way through the test, from their off-peak schedule to their afternoon-traffic schedule, while in Singapore lights are adaptive, using detectors embedded under the road to determine how much traffic is around and thus when a signal should change. In neither case was SignalGuru fooled. Fuel consumption fell, too—by about 20%. SignalGuru thus reduces both frustration and fuel use, and makes commuting a slightly less horrible experience.