this paper presents an application of advanced signal processing techniques in the determination of transportation network critical routes for control purposes. The proposed method can be considered as an alternative to the orthodox O-D estimation methods. During the peak periods, certain movements in the network are considered significant based on their location (i.e., entry and exit points) and flow fluctuation. The proposed method considers observed counts profiles resulting from the peak demand and signal operation as non-stationary time series. Wavelet domain processing is used to decompose, de-noise, compress, and extract the common patterns in a set of traffic flow time series collected from several detectors in signalized sub-network in Reston Parkway in Northern Virginia. The results show that several matched patterns in movements can be detected. The obtained route''s pattern matches the field observation. The proposed method was found viable in identification of critical routes under congested conditions‥
展开▼
