The traffic prediction and special factor adjustment model using artificial intelligence

The research has presented a heuristic model which can predict link travel speed and its variations under the influence of special factors, including incident and construction.; This dissertation addresses a new framework of short-term traffic prediction model, which combines technologies of artificial intelligence such as wavelet transform, fuzzy logic and neural network. In the mean time, the model proposed in this dissertation is also used to quantitatively calculate the influences of special factors.; First, the wavelet de-noising method is used to extract the standard traffic patterns based on historical data. Then, the neuro-fuzzy network which implements fuzzy logic inside the neural network is employed to identify the influences of special factors. Finally, the traffic prediction is conducted by applying the effects of special factors to the standard pattern. The prediction models are calibrated using historical data and the results have demonstrated promising performance. Subsequently, recommendations are made regarding future research in this area.