Abstract:Recently, a phase transition to synchronized congested traffic has been observed in empirical highway data [B. S. Kerner and H. Rehborn, Phys. Rev. Lett. 79, 4030 (1997)]. This hysteretic transition has been described by a non-local, gas-kinetic-based traffic model [D. Helbing and M. Treiber, Phys. Rev. Lett. 81, 3042 (1998)] that, however, did not display the wide scattering of synchronized states. Here, it is shown that the latter can be reproduced by a mixture of different vehicle types like cars and trucks. The simulation results are in good agreement with Dutch highway data.
What problem does this paper attempt to address?
The problem that this paper attempts to solve is: How to reproduce the widespread dispersion phenomenon of traffic flow data in the congested state through macroscopic simulation, especially in the synchronous congestion state. Specifically, the authors hope to explain why in the synchronous congestion state, the data points of traffic flow and density are distributed in a two - dimensional area, rather than following a single one - dimensional relationship as in the free - flow state.
### Main Problem Background
1. **Synchronous Congestion Phenomenon**:
- Recent studies have observed a phase transition from the free - flow to the synchronous congestion state in the measured data on highways (Kerner and Rehborn, 1997). This phase transition has hysteresis and usually occurs at ramps, slopes or other bottlenecks.
- Previous studies used traffic models based on non - local gas dynamics to describe this phenomenon (Helbing and Treiber, 1998), but failed to reproduce the widespread dispersion of data in the synchronous congestion state.
2. **Reasons for Data Dispersion**:
- In the free - flow state, the flow - density graph can be well described by a unique one - dimensional relationship.
- However, in the congested state, the measured data points are distributed in a two - dimensional area, indicating significant fluctuations and irregularities.
3. **Existing Explanations**:
- Krauß proposed that driver behavior is different in the congested state from that in the free - flow state.
- Lenz et al. and Schreckenberg believe that this dispersion is caused by the anticipatory effect of drivers on the vehicles ahead and the further traffic dynamics.
### Main Contributions of the Paper
The authors propose that this widespread dispersion phenomenon can be explained by introducing different types of vehicles (such as cars and trucks). The specific methods are as follows:
- **Influence of Heterogeneous Traffic Groups**: Different types of vehicles (e.g., cars and trucks) have different parameter sets, which define the equilibrium flow - density relationships for pure car traffic and pure truck traffic. In mixed traffic, by taking a weighted average of these parameter sets, the heterogeneity in actual traffic can be simulated.
- **Introduction of Boundary Conditions**: By introducing the measured boundary conditions (including the proportion of trucks in different time periods), the widespread dispersion phenomenon in the synchronous congestion state can be reproduced in the macroscopic simulation.
- **Simulation Results**: The authors used the actual data on the A9 highway in the Netherlands for verification and found that the simulation results are in good agreement with the measured data.
### Key Formulas
1. **Continuity Equation**:
\[
\frac{\partial \rho}{\partial t}+\frac{\partial (\rho V)}{\partial x}=\frac{Q_{rmp}}{nL}
\]
where \( \rho(x,t) \) is the vehicle density, \( V(x,t) \) is the average vehicle speed, \( Q_{rmp} \) is the inflow or outflow volume of the ramp, \( n \) is the number of lanes, and \( L \) is the merging length.
2. **Average Velocity Evolution Equation**:
\[
\frac{\partial V}{\partial t}+V\frac{\partial V}{\partial x}=-\frac{1}{\rho}\frac{\partial (\rho \theta)}{\partial x}+\frac{1}{\tau}(U - V)
\]
where \( \theta \) is the velocity variance, \( U \) is the dynamic equilibrium speed, and \( \tau \) is the relaxation time.
3. **Dynamic Equilibrium Velocity Formula**:
\[
U = V_0\left[1-\frac{\theta+\theta'}{2A(\rho_{max})}\left(\frac{\rho'}{1-\rho'/\rho_{max}}\right)^2B(\delta V)\right]
\]
4. **Dynamics Approximation of Variance**: