Distributed Identification Of The Cell Transmission Traffic Model: A Case Study.pdf

## Delft University of Technology - Distributed Identification of Cell Transmission Traffic Model

This technical report from Delft University of Technology's Delft Center for Systems and Control presents a distributed identification method for the Cell Transmission Model (CTM), a first-order macroscopic traffic model describing vehicle density in different sections (cells) of a freeway.

Key Points:

Problem: Identify parameters (free flow speed, maximum density, backward congestion propagation speed) of the CTM to predict and control traffic flow accurately, especially during events like accidents or weather changes. Traditional methods are computationally intensive for real-time applications.

Approach: Proposes a distributed identification scheme that:
Minimizes prediction error of cell densities one step ahead.
Reduces computation time and effort compared to centralized identification.

Model Validation: The methodology is validated using real-life data from a portion of the A12 freeway in the Netherlands.

Performance Evaluation: The identified model's performance is assessed under various scenarios, demonstrating its effectiveness for virtual sensor applications.

CTM Overview:

The CTM models traffic flow as a network of interconnected cells, each with:

Density (ρi(k)): Number of vehicles per lane per kilometer at time step k.
Flow (qi(k)): Volume of traffic leaving the cell per hour at time step k.

Cells are linked by junctions with on-ramps and off-ramps, influencing flow between them. The model incorporates fundamental traffic dynamics through equations describing density evolution and flow regulation at each cell.

Main Contributions:

A distributed identification scheme for the CTM parameters that leverages hierarchical and distributed optimization techniques.
Validation using real data and performance evaluation under diverse conditions.

Significance:

This work advances the field of traffic modeling and control by providing a computationally efficient method for identifying CTM parameters crucial for accurate prediction, monitoring, and control in dynamic traffic scenarios.