Model-Based Approach for Queue and Delay Estimation at Signalized Intersections with Erroneous Automated Data

A model-based scheme has been developed in this paper to estimate the number of vehicles in queue and the total delay, which are typical measures for characterizing a signalized intersection. The estimation was first carried out by using the input-output and queue accumulation polygon methods. These were evaluated by using data from two instrumented intersections in the city of Lincoln, Nebraska. However, the estimates did not agree well with the actual data primarily because of the errors in counts obtained from the automated detectors. Hence, an analysis of the characteristics of the errors was carried out. A model-based estimation approach using the Kalman filter was then developed. Suitable modifications were further applied to the estimation process by incorporating calibration constants for particular site/traffic conditions in the Kalman filter estimation scheme. The results obtained were promising, indicating that the scheme could be used for performance analysis of signalized intersections with automated detectors. (C) 2016 American Society of Civil Engineers.