|
A FCD–based Method for Traffic Congestion Detection in low-proximity data environment |
|---|---|
| รหัสดีโอไอ | |
| Creator | Ilya S. Ardakani |
| Title | A FCD–based Method for Traffic Congestion Detection in low-proximity data environment |
| Contributor | Thanaruk Theeramunkong, Waree Kongprawechnon |
| Publisher | Sirindhorn International Institute of Technology, Bangkadi Campus (SIIT-BKD) |
| Publication Year | 2560 |
| Journal Title | Journal of Intelligent Informatics and Smart Technology |
| Journal Vol. | 1 |
| Page no. | 1-7 |
| Keyword | Intelligent Transportation Systems, Artificial Neural Networks, Pattern Recognition Traffic Congestion Detection, Expectation Maximization, Gaussian Mixture Models |
| URL Website | https://ph05.tci-thaijo.org/index.php/JIIST |
| Website title | Journal of Intelligent Informatics and Smart Technology |
| ISSN | 2586-9167 |
| Abstract | Recently GPS data gathered from vehicles on roads, so-called floating car data (FCD), have become one of the popular sources for traffic management. As this source of data becomes more accessible, there are increasing research works conducted in order to extract reliable information from available raw sensory data. Such data from public transportation equipped with GPS sensors provides relatively consistent source of FCD. However, in some cases, due to low-frequency travel routines and the small number of active vehicles, extracted FCD are not sufficient for traffic prediction. This study proposes a method to efficiently identify road congestion using FCD collected from in-campus shuttle buses in low speed routes. In this task, the challenging issue is how to distinguish real congestion from intentional delay, such as passenger loading and temporary parking, since both are similar in terms of low average speed on the road link. The proposed method employs artificial neural network to predict traffic on the links with sparse FCD available. The experimental result demonstrates that our method is capable of producing comparable results with existing methods. |
