Estimating Vehicular Traffic Count Using Mobile Phone Network for Effective Traffic Management

Abstract

Obtaining quality information about traffic congestion for effective traffic management of a location requires an expensive means with tedious operations. Generally, existing methods such as the use of manual counting, global positioning system (GPS), radar, inductive loops are all costly with respect to time, personnel and complexity. In this paper, mobile phone flow is being measured instead of measuring vehicle flow (traffic flow). We made use of a k-means clustering technique on a network-triggered data to estimate the vehicular traffic count of a highway. Results of this technique was evaluated with the official count and gives a mean percentage error of 12.96%. This technique proves to be reliable and cost-effective in estimating traffic count which could be used on a large scale.