| Traffic flow state recognition is one of the important issues for Intelligent Transportation Systems research, especially for Advanced Traffic Management Systems and Advanced Traveler Information Systems research. It is a significant precondition to establish derivational and control measures correctly by the research on traffic flow state and the recognition on traffic abnormity timely and accurately. Because of the limitation of Mathematics, the former research is mainly concentrated on traffic flow state recognition in advance (namely traffic flow forecasting) and real-time recognition (namely incident detection or traffic flow qualitative change detection), but recognition of quantitative change rule of traffic flow is neglected. The quantitative change is the precondition and essential provision for qualitative change, and the latter is the certain results for the accumulation of the former. The change of traffic parameters begin with quantitative change, furthermore, there must be a qualitative change with the increase of quantitative change. Accordingly, the study on quantitative change rule makes a significant sense for the real time recognition of traffic flow state.Based on the reviews of related literatures both in and abroad, the quantitative change of traffic flow state has been recognized accurately by the analysis of change-point of traffic flow models mentioned below, that is, the study on the recognition of change-point of traffic flow linear regressions model based on projective transform, the study on the recognition of change-point of traffic flow linear regressions model based on unitary regression method, the study on the recognition of change-point of multiple linear regressions model, the study on the recognition of change-point of non-linear regressions model, the study on the recognition of change-point of time series model including both stationary and nonstationary ones. These methods are calibrated and tested with the field data of typical road of Zibo city and Southampton, U K to verify the validity and the feasibility of the theory. |
