Macroscopic 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 data and flow estimated using the proposed methodology. The results showed that with the proposed scheme, average error in estimated density for the sections under consideration was around 20.8%, whereas it was found to be 30% with direct use of automated sensor data. This shows that at error prone locations, the proposed model is a viable alternative to the direct use of erroneous automated data.
Traffic flow estimation at error prone locations using dynamic traffic flow modeling
Transportation Letters ; 11 , 1 ; 43-53
2019-01-02
11 pages
Aufsatz (Zeitschrift)
Elektronische Ressource
Englisch
Traffic flow estimation at error prone locations using dynamic traffic flow modeling
Online Contents | 2017
|Identification of Network Sensor Locations for Estimation of Traffic Flow
Transportation Research Record | 2014
|Europäisches Patentamt | 2023
|Europäisches Patentamt | 2020
|Europäisches Patentamt | 2022
|