Traffic Flow Prediction Using Time-Series and Autocorrelation Criterion

To develop a traffic management system we must first implement a system identifier that can use input data which have been taken from sensors installed inthe routes. The output of the system identifier would be the predicted traffic flow based on the past and/or present traffic flow. Having a viewpoint of the future oftraffic congestion would be a good help to manage the traffic. Other implementations would be beneficiary for this system to upgrade the usability and applicability but in this paper we focus on the prediction part of the system.

Keywords:

Traffic management , Neural Network , Time- Series , Multi-Layer Perceptron , Levenberg- Marquardt Algorithm