Novel multi input parameter time delay neural network model for traffic flow prediction

Intelligent Transportation Systems are an important aspect of our life and are going to become ubiquitous in the near future. Traffic flow prediction is a key component of any Intelligent Transportation Systems. This report uses Artificial Neural Network based models to predict short term traffic flow. Two new input parameters; temperature and truck flow has been introduced into a multi input parameters model. Accordingly five models have been developed based on different combination of the input parameters. By considering past values of multiple input parameters as compared to past values of traffic flow only, the percentage reduction in RMSE values ranges from four to nineteen percentages. Hence, it is suggested to use multi input parameter models for predicting short term traffic flow.