| Summary: | For the past few years, as the Intelligent Transportation System (ITS) developing rapidly, intelligent transportation control and management has become a popular topic. In many countries, many people rely on the public transport system for commuting. Commuters concern more about the reliability and punctuality of the public transport system. Therefore, the precise prediction of real-time traffic conditions has become the key of the transport management system. As is well-known, road traffic system is human-related, time-varying and complex massive system. It has high uncertainty, due to natural factors (season and weather) and artificial reasons (traffic accident and drivers’ mentality). These factors bring more challenges to the prediction of traffic flow, especially for short-term forecast. This thesis works on the short-term prediction which is different from macroscopic aspect. The approach is with the help of machine learning utilizing dynamic neural network.
|
|---|