01660943

Component

Technological advances allow the identification of vehicles’ passage at given locations and at low investment costs. These technologies collect data only for an unknown sample of vehicles from the entire population. Therefore, aggregated measures such as flow or density cannot be inferred from them. Since these data are relevant for evaluation purposes, the deployment of usually expensive infrastructure-based sensors is required. It seems appealing to find a method to estimate the detection rate from the detections themselves without the need for more expensive measurements. In this paper, the authors propose a heuristic method to estimate the detection rate of virtually any faulty vehicle-detection technique. The method starts from a known series of times between detections, assuming a single lane, a constant detection rate and a family of probability distributions for the underlaying headways. The method is validated using simulated detections over real headway data. The method’s accuracy depends on the assumed family of distributions and the amount of detections, working best for higher detection rates and/or lengths of the detection period. The method seems to slightly overestimate the detection rate. The authors discuss a number of alternatives to alleviate this issue as part of both this and future research.

This paper was sponsored by TRB committee AHB15 Standing Committee on Intelligent Transportation Systems.

18-05734

English

13p

2018

Transportation Research Board 97th Annual Meeting

Digital/other

Figures; References

Automatic vehicle detection and identification systems; Bluetooth technology; Distributions (Statistics); Estimating; Flaw detection; Headways; Heuristic methods; Simulation; Time intervals; Traffic data; Traffic flow; Vehicle detectors

Data and Information Technology; Highways; Operations and Traffic Management; Vehicles and Equipment

Transportation Research Board Annual Meeting 2018 Paper #18-05734

TRIS, TRB, ATRI

Jan 8 2018 11:28AM