Spatiotemporally Discrete Urban Traffic Model.pdf

Spatiotemporally Discrete Urban Traffic Model

This technical report from Delft University of Technology's Center for Systems and Control presents a spatiotemporally discrete urban traffic model designed for model-based traffic control.

Key Features:

Macro-scale Approach: The model focuses on macroscopic traffic dynamics, representing average vehicle behavior rather than individual vehicle details. This trade-off balances modeling accuracy with computational efficiency.
Variable Sampling: Unlike traditional models with fixed sampling intervals, this approach allows for variable sampling time intervals and space distances. This flexibility enables optimization of both modeling accuracy and computation speed based on specific control needs.
CFL Condition: The report introduces an urban traffic CFL (Courant-Friedrichs-Lewy) condition to ensure the discrete model accurately captures necessary traffic phenomena.

Model Structure:

The model discretizes a continuous queue-based first-order traffic flow model into spatially and temporally sampled segments:

Links: Represent road segments between intersections controlled by traffic signals (see Figure 1).
Vehicles: Queues form at intersection approaches, with vehicle arrival rates and departure rates modeled based on signal phasing.
Sampling:
Time: Sampling intervals vary dynamically based on traffic conditions to optimize computational efficiency.
Space: Road segments are discretized into fixed spatial distances for consistent representation across the network.

Advantages:

Versatility: The model can be adapted to various traffic scenarios (unsaturated, saturated, over-saturated) by adjusting sampling parameters.
Computational Efficiency: Variable sampling intervals allow for faster simulation compared to models with fixed intervals while maintaining acceptable modeling accuracy.
Control Applicability: The model's balance between accuracy and speed makes it suitable for real-time traffic control applications.

Future Research:

The report highlights the need for further investigation into:

Optimal sampling strategies for specific traffic conditions and control objectives.
* Validation of the model against real-world traffic data.