Short-term traffic flow prediction using different techniques

Due to the complexity of traffic flow characteristics and the drawbacks of the traditional methods, the short-term predictions using the existing individual methods generally lack accuracy and robustness during all the time periods of the day. In order to overcome the drawbacks of traditional methods, the present paper proposes a fuzzy rule-based system (FRBS), which is used to combine the traffic flow forecasts resulting from the Exponential Smoothing Method (ESM), Autoregressive Integrated Moving Average (ARIMA), Artificial Neural Networks (ANN) and the Fuzzy Logic System (FLS). A comparative study shows that the proposed FRBS can represent the traffic flow more accurately and the results gained from the FRBS are found to outperform those of the traditional methods, when modelling in the real urban traffic.