New road construction in order to reduce congestion is almost impossible in developed countries, due to the lack of budget and to the difficulty of securing land for the roads. To make the best use of the existing roads, proactive traffic control strategies and traffic demand management are needed. These must be based on predictive traffic information.
These forecasting capabilities are achieved using long term and short term data. Long term data refers to the historical measurements, and it is useful to predict recurrent congestion. Non-recurrent congestion needs to be predicted by using short term data. A methodology to make reliable predictions up to one hour using these types of information will be presented.
The methodology is applied to a case study in Seoul.