Simple analytical models for estimating the queue lengths from probe vehicles at traffic signals: A combinatorial approach for nonparametric models.

This study develops a combinatorial approach for nonparametric short-term queue length estimation in terms of cycle-by-cycle partially observed queues from probe vehicles (PV). The method does not assume random arrivals and does not assume any primary parameters or estimation of any parameters but uses simple algebraic expressions that only depend on signal timing. For an approach lane at a traffic intersection, the conditional queue lengths given probe vehicle location, count, time, and analysis interval (e.g., at the end of the red signal phase) are represented by a Negative Hypergeometric distribution. The simple analytical estimators obtained are compared with parametric methods from literature and highway capacity manual methods using field test data and simulation data involving probe vehicles. The analysis indicates that the nonparametric models presented in this paper match the accuracy of the parametric ones used in the field test and simulated data for estimating queue lengths. • Two new models are derived for short-term queue length estimation. • The models are based on combinatorics and nonparametric. • The models are closed-form input–output formulas. • New probability mass functions for traffic queues are derived. • The models can be used in connected traffic signals. [ABSTRACT FROM AUTHOR]