自動学習機能を備えた高速道路旅行時間予測モデルの開発 Online-learning Type of Traveling Time Prediction Model in Expressway
In expressway, there exist a lot of requirements for traveling time prediction as the information on express way traffic-flow. It is a problem to be solved to predict traveling time with high accuracy in case traffic-flow changes dynamically like the beginning or the end of traffic congestion. It is therefore important to make the trafficsimulation model using time-series data by vehicle detectors. In this case, it is necessary to cope with secular change of traffic-flow characteristics like constructions of new expressways or environment condition changes. Thispaper proposes the new traffic-flow prediction system which has the model learning function by time-series data processing. The system makes the relation model between traffic density and vehicle velocity using neural networks. The neural networks is used as on-line learning model VENN(Vehicle velocity Estimation model on Neural Network). The system simulates the traffic-flow using the VENN and predicts traveling time using the simulation results. The proposed system was already estimated in the actual expressway, which ended in satisfactory results.
- 電気学会論文誌. D, 産業応用部門誌 = The transactions of the Institute of Electrical Engineers of Japan. D, A publication of Industry Applications Society
電気学会論文誌. D, 産業応用部門誌 = The transactions of the Institute of Electrical Engineers of Japan. D, A publication of Industry Applications Society 117(8), 980-985, 1997-08
The Institute of Electrical Engineers of Japan