Traffic flow estimation at error prone locations using dynamic traffic flow modeling

AbstractMacroscopic model-based schemes are appropriate for real-time estimation of traffic density, which is an important congestion indicator. Conservation of vehicles equation is one of the basic equations used in any macroscopic model-based analysis. However, to apply the conservation equation, accurate flow/count data should be available, which can be achieved only if suitable traffic sensors are available. In reality, automated traffic sensors are prone to measurement errors, especially under heterogeneous, less lane-disciplined and congested traffic conditions. Use of this erroneous data can lead to wrong estimation of traffic variables. This study proposes a lumped parameter macroscopic model-based scheme for the estimation of flow at error-prone locations. The estimation scheme was designed based on the extended Kalman filter. The performance of the proposed scheme was evaluated for density estimation by comparing the results using accurate flow data collected manually, using automated sensor dat...