Incorporating neural network traffic prediction into freeway incident detection

The efficient operation of an incident management system depend Neural network models have been applied to traffic prediction frequently and even repeatedly because of its superior capability in emulating nonlinear systems. However, these traffic prediction models have not been utilized for incident detection. On the other hand, it is expected that the performance of an incident detection algorithm can be improved if an advanced prediction model is incorporated into. Therefore, this study developed several traffic prediction models that were then integrated into incident detection algorithms. The traffic prediction models were developed based on three different choices of independent variables, while the incident detection algorithms employed different decision functions. The results show that a good prediction model can improve the performance of an incident detection algorithm only when the decision function of the algorithm is appropriately chosen