Research On The Discrete Modeling Method Within The Fundamental Diagram Approach And Three Phase Traffic Flow Theory

Engineering practice proves that the first principle of relieving congestion is investigating and exploring the mechanisms of traffic flow. Traffic flow theory was developed under this background, whose main branches are the traditional fundamental diagram approach and the novel three phase traffic flow theory. Fundamental diagram approach assumes the unique relationship between flow and density, while three phase traffic flow theory denies this assumption. The controversies between the fundamental diagram approach and the three phase traffic flow theory are one of the hotspots in the traffic flow theory. This dissertation is based on these theories:1) Based on the Nagatani lattice hydrodynamic model, the density difference lattice model (DDLM) is proposed. DDLM not only considered the leading lattice but also the following lattice. Linear stability curve are obtained by the linear stability analysis, which demonstrates that the density difference can enhance the stability. Burgers, KdV, mKdV equations and the coexistence curve are derived by the nonlinear stability analysis. Simulations show that the triangular shock wave, soliton and kink and anti-kink wave described by these equations are simulated. Then the lattice hydrodynamic model with the onramp system is proposed and the stochastic and deterministic onramp simulation methods are designed. Onramp simulations are in accordance with the empirical findings. Moreover four new kinds of congested patterns are founded, two of which are very similar to the GP and Dissolving GP in the three phase theory. Thus DDLM not only confirms the empirical findings of the fundamental diagram approach, but also put forward new findings which need to be consolidated and proves that wide moving jams may not emerges in the free flow in the models of fundamental diagram approach.2) The synchronized flow of the brake light cellular automaton model (BLM) and the traffic breakdown of FMCD and BLDAD brake light cellular automaton models are not in accordance with the real traffic. Based on BLM, the Desired Time Gap Brake Light Model (DTGBLM) is established by incorporating the desired time gap. Simulations show that the synchronized flow of DTGBLM is in accordance with the empirical results, such as the WSP and GP founded by the three phase traffic flow theory. Deceleration analysis demonstrates that although there are unrealistic decelerations, its occurrence probability is very small and can be neglected. Based on FMCD and BLDAD, the Advance Randomization Brake Light Model (ARBLM) is proposed. The mechanisms to reproduce the synchronized flow and related traffic breakdown phenomenon of ARBLM are investigated. All results show that the criticisms of break light model that the three phase theory imposed are solved by DTGBLM and ARBLM.3) Combining the classical car following model which considers of multiple leading vehicles with the NaSch model, the Average Space-Gaps cellular automaton Model (ASGM) within the fundamental diagram approach is proposed. Simulations show that the ASGM produces the same spatiotemporal dynamics as many of the more complex three-phase models. Besides many aspects that are consistent with traffic data. It seems that the absence of a unique fundamental diagram is not an essential requirement for this model class, at least with respect to the spatiotemporal dynamics. With the further consideration of the classical car following model which considers of multiple leading vehicles, the Extended ASGM (EASGM) is proposed, which can reproduce the synchronized outflow of wide moving jams. ASGM and EASGM not only bridge the cellular automaton models with car following models, but also invalidate the criticism that models within the fundamental diagram approach cannot reproduce the synchronized outflow of wide moving jams.4) In order to verify the validity of the proposed models, some models under different background are calibrated and validated with the new models proposed in this paper. The three stage calibration method is explored. In the first stage, the parameters that can be determined by the default or empirical values are picked out. In the second stage, other parameters are calibrated by the traditional calibration methods. In the last stage, one should calibrate the parameters that influence the velocity directly, which are chosen from the parameters in the first stage by traditional calibration methods. Calibration results show that all proposed models in this paper are better than their original models. And models in the three phase traffic flow theory or in fundamental diagram approach has nothing to do with the calibration results.