A single loop detector-based freeway speed estimation

This paper aims to develop a freeway speed estimation algorithm using single-loop detector. To achieve this, a quadratic mathematical relationship describing the behavior of traffic variables such as speed, flow and occupancy has been reviewed and formulated. We concluded that the occupancy could be expressed as quadratic linear combinations of flow and/or speed, and thereby the speed could be calculated using the quadratic equation calibrated. The calibration has been performed using the field traffic data to obtain the parameters of the quadratic regression equation. They have been determined at each detector station, where the measured speed information is available using the double-loop detector. Using the calibrated coefficients and the traffic data for each station, the speed estimation has also been performed. To evaluate the fitness of the model the MAE (mean absolute errors) indicator has been employed. The average MAE of speed over the surveyed spatial scope has been identified as only 3.7 km/h, meaning that the proposed model seems to be quite satisfactory and could be applicable to actual usages. Some issues of transferability of the calibrated parameters, in terms of both temporal and spatial aspects, have also been discussed along with some future research agenda.