Application of time-series and wavelet analysis for traffic flow modelling and incident detection

Abstract

Efficient and sustainable use of existing transportation network has become a major issue in socio-economic growth of the developed parts of today’s world. Intelligent Transportation System (ITS) is an important tool to increase the sustainability and efficiency of an existing traffic management system. Implementation of ITS in an existing transport network requires the use of efficient short-term traffic condition (such as, traffic flow, density or speed) forecasting and incident detection algorithms. In the present study, univariate and multivariate traffic flow modelling and incident detection techniques applicable to ITS equipped urban traffic control (UTC) systems are proposed. The proposed models are validated at single or multiple junctions within the congested urban transport network at the city-centre of Dublin.