Dynamic Classification of Traffic Flow Patterns Simulated by a Switching Multimode Discrete Cell Transmission Model

In this paper, a dynamic approach to specify flow pattern variations simulated by a multimode macroscopic flow model is followed, incorporating the neural network theory to reconstruct real-time traffic dynamics. In order to deal with the noise in and the wide scatter of traffic data, filtering is applied prior to overall modeling process. Filtered data are dynamically and simultaneously input to neural density estimation and traffic flow modeling processes. Traffic flow is simulated by modifying the cell transmission model in order to explicitly account for flow condition transitions considering wave propagations. Cell-specific flow dynamics are used to determine the mode of prevailing traffic conditions, which are, in turn, sought to be reconstructed by neural methods. The classification of flow patterns over the fundamental diagram is obtained by considering traffic density as a pattern indicator. The fundamental diagram of speed-density is updated to specify the current corresponding flow pattern. The modified classification returned promising results in capturing sudden changes on test stretch flow patterns that are simulated by the switching multimode discrete macroscopic model.