iTraffic: A Smartphone-based Traffic Information System

Abstract

We propose the i-Traffic system that utilizes crowdsourced data from smartphones for the traffic flow mining by shockwave techniques. Shockwave is the propagation phenomenon of vehicle accumulation or relief on roads between two traffic flows with different speeds. The movement data of vehicles in front of an intersection are collected via smartphones for the shockwave identification. To conquer the low penetration problem when the number of the movement data is low, a folding heuristic is proposed by using traffic light cycle information to virtually increase the penetration of movement data. We implement our system on a client-server architecture and perform a small scale field trial experiment to demonstrate the system capability. Our results showed that our system is able to compute traffic information, including red/green light transition information and vehicle arrival rate with mean absolute errors of 5.0/0.6 seconds and 2.43 vehicles per minute, respectively under a low penetration rate of 1.2%.