Abstract:We present a macroscopic model of mixed multi-lane freeway traffic that can be easily calibrated to empirical traffic data, as is shown for Dutch highway data. The model is derived from a gas-kinetic level of description, including effects of vehicular space requirements and velocity correlations between successive vehicles. We also give a derivation of the lane-changing rates. The resulting dynamic velocity equations contain non-local and anisotropic interaction terms which allow a robust and efficient numerical simulation of multi-lane traffic. As demonstrated by various examples, this facilitates the investigation of synchronization patterns among lanes and effects of on-ramps, off-ramps, lane closures, or accidents.
What problem does this paper attempt to address?
This paper attempts to solve several key problems in multi - lane traffic flow, specifically including:
1. **Macro - modeling of multi - lane traffic flow**: The author proposes a macro - model capable of describing multi - lane mixed traffic flow. This model is based on gas - kinetic description and takes into account the spatial requirements of vehicles and the correlation between the speeds of the front and rear vehicles.
2. **Modeling of lane - changing behavior**: The paper not only considers the interactions between vehicles (such as overtaking or avoiding interactions with slow - moving vehicles), but also introduces "spontaneous" lane - changing behavior (reflecting traffic regulations and drivers' personal preferences) and "forced" lane - changing (such as lane merging, on - and off - ramps, etc.). These factors jointly affect the synchronization patterns between lanes and traffic flow characteristics.
3. **Non - local and anisotropic interaction terms**: By introducing non - local and anisotropic interaction terms, the model can more accurately simulate actual traffic conditions, especially when dealing with unstable multi - lane traffic problems. These interaction terms allow for robust and efficient numerical simulations of traffic flow.
4. **Model calibration and validation**: The author shows how to calibrate the model to actual traffic data (for example, data from the A9 highway in the Netherlands) to ensure the practicality and accuracy of the model.
5. **Research on complex traffic phenomena**: Using the proposed model, various complex traffic phenomena can be studied, such as synchronization patterns between lanes, the influence of on - and off - ramps, traffic changes caused by lane closures or accidents, etc.
In summary, this paper aims to develop a macro - model that can accurately describe multi - lane traffic flow and improve existing models by introducing detailed lane - changing behaviors and non - local interaction terms, thereby better understanding and predicting complex traffic dynamics. This is helpful for the development of traffic management and optimization measures.
### Key formulas
- **Continuity equation of phase - space density**:
\[
\frac{\partial f_a^i}{\partial t}+\frac{\partial (f_a^i v)}{\partial x}+\frac{\partial \left( f_a^i \frac{V_{0a}^i - v}{\tau_a^i} \right)}{\partial v}=\left( \frac{\partial f_a^i}{\partial t} \right)_{\text{int}}+\left( \frac{\partial f_a^i}{\partial t} \right)_{\text{lc}}
\]
- **Interaction rate**:
\[
I_{ab}^i=\chi_i\int_{v > w}dw (v - w) f_{ab}^i(x, v, x + s_a^i, w, t)
\]
\[
J_{ab}^i=\chi_i\int_{w > v}dw (w - v) f_{ab}^i(x, w, x + s_a^i, v, t)
\]
- **Boltzmann factor**:
\[
A(\delta V)=\chi(\rho)\rho\rho'\sqrt{S}[N(\delta V)+\delta V E(\delta V)]
\]
\[
B(\delta V)=\chi(\rho)\rho\rho'S\left[ \delta V N(\delta V)+(1 + \delta V^2)E(\delta V) \right]
\]
\[
C(\delta V)=\chi(\rho)\rho\rho'S\left[ V\sqrt{S}N(\delta V)+(\theta - k\sqrt{\theta\theta'}/S + V\sqrt{S}\delta V)E(\delta V) \right]
\]
These formulas are used to describe the interactions between vehicles and their effects on traffic flow dynamics.