01697447

Component

Using probe vehicle data to estimate maximum queue length for signalized intersections is recently of great interest to both researchers and practitioners. The problem becomes more challenging when the available probe vehicles are in small size. Many efforts have been devoted for queue length estimation in such a situation, yet it’s still unready for practical application. The main reason is that those studies either rely on strong assumptions or have high variance. In this paper, the authors propose a Bayesian algorithm that manages various practical issues, including insufficient samples, noise data, random and time-varying traffic, real-time signal control, etc. The queue length is inferred by the maximum a posteriori (MAP) method through a few of techniques: jointly use probe vehicles from adjacent cycles, novelly develop tight upper bounds of queue length; carefully filter noise; and naturally embed past information. Simulation experiments have confirmed the superiority of the authors' algorithm by comparison of two benchmark algorithms. The robustness and accuracy of the algorithm is further demonstrated by a case study with sensitivity analysis.

This paper was sponsored by TRB committee ABJ70 Standing Committee on Artificial Intelligence and Advanced Computing Applications.

19-02009

English

22p

2019

Transportation Research Board 98th Annual Meeting

Digital/other

Appendices; Figures; References; Tables

Data processing operations; Probe vehicles; Queuing; Signalized intersections; Traffic measurement

Bayesian statistics

Data and Information Technology; Highways; Operations and Traffic Management

Transportation Research Board Annual Meeting 2019 Paper #19-02009

TRIS, TRB, ATRI

Dec 7 2018 9:27AM