Clearance Estimation through Mobile Sensing

Abstract

Clearance is a spatial gap between two successive vehicles, which can be used to indicate the traffic condition. Obviously, the average clearance in a congested traffic will be smaller than that in a free-flow traffic. In most current infrastructure-based traffic information systems, clearances between vehicles could be estimated from an image or a video captured by fixed sensors such as traffic surveillance cameras. However, using fixed sensors is not effective in terms of cost, coverage, and convenience. A mobile sensing approach, which vehicles act as mobile sensors and collect the traffic data as they travel, is more appealing. In this paper, we explore a possibility of using a built-in camera on a smartphone for clearance estimation. The estimation algorithm and its accuracy will be discussed. Particularly, our experimental results show that the mean absolute errors of the estimates are within 10% of the actual reference values. This is suitable for a traffic sensing application.