Intelligent Transport Systems

10 Congestion Prediction by Means of Fuzzy Logic and Genetic Algorithms

Xiao Zhang¹, Enrique Onieva² Victor C.S. Lee¹ and Kai Liu³

¹ City University of Hong Kong, Hong Kong

² University of Deusto, Bilbao, Spain

³ Chongqing University, China

10.1 Introduction

Traffic congestion causes energy waste, requires investment in public infrastructure, and threatens urban environmental quality. These challenges are even more pressing if the forecast of growth in transport is considered. The business costs caused by congestion will increase by approximately 50% by 2050, according to the Transport White Paper (European Commission, March 2011). Intelligent Transportation Systems (ITS) consider easing highway traffic congestion a significant issue to be investigated. Therefore, the prediction and identification of traffic congestion play a vital part in ITS, which intends to make it more efficient, safer and energetically sustainable. Accurate traffic prediction reporting can be adopted either by drivers to avoid traffic jams, or traffic management systems to take measures in advance to ensure traffic flow.

In the last few decades, the most commonly used techniques for traffic forecasting are on the basis of the Kalman Filter (KF) [1,2] and the Autoregressive Integrated Moving Average (ARIMA) [3,4]. Although these techniques can obtain good results, there are still some weaknesses. For example, KF tends to generate overestimation or underestimation that deteriorates the prediction ...