Traffic Condition Estimation Based on Historical Data Analysis

Traffic condition estimation is one of the essential tasks for intelligent transportation systems/services (ITS). In order to provide accurate and timely traffic information, traffic data must be collected adequately at every location or road segment in the traffic network. This requirement is hard to satisfy, however, in practice considering a large Spatio-temporal space of urban traffic system even both fixed-sensor and crowd-sourcing approaches are utilized. In order to resolve this difficulty, this work proposes a framework for analyzing historical data to predict traffic conditions at the road segments where real-time data are missed by leveraging data mining techniques. The experimental results from real-field data collected by our developed system reveal the feasibility and the effectiveness of the proposed approach.