Freeway Travel Time Prediction Research Based on A Deep Learning Approach
Junfeng Zhang, Hongxi Chen, Hong Zhou, Zhihai Wu
Available Online September 2016.
- https://doi.org/10.2991/amitp-16.2016.97How to use a DOI?
- Deep learning, Toll data, Travel time prediction, Automatic encoder
- Accurate prediction of freeway travel time is quite significant for traffic management and travelers' trip decisions. A model of stacked automatic encoders for freeway travel time prediction is proposed. First, obtain the hourly average travel time by processing freeway toll data. Considering data quantity and travel time scheduling, a travel time prediction model based on stacked automatic encoders is established then. The automatic encoder is trained with unsupervised learning step by step by using historical travel time data, adjust the parameters of prediction layer with gradient descent method. Ultimately the one-hour travel time is predicted using the past three-hours' average travel time. Compared to the traditional BP neural network prediction model, the Root Mean Square Error of this method drops by 13.6% ,verifying the validity of the model.
- Open Access
- This is an open access article distributed under the CC BY-NC license.
Cite this article
TY - CONF AU - Junfeng Zhang AU - Hongxi Chen AU - Hong Zhou AU - Zhihai Wu PY - 2016/09 DA - 2016/09 TI - Freeway Travel Time Prediction Research Based on A Deep Learning Approach BT - Proceedings of the 2016 4th International Conference on Advanced Materials and Information Technology Processing (AMITP 2016) PB - Atlantis Press SP - 487 EP - 493 SN - 2352-538X UR - https://doi.org/10.2991/amitp-16.2016.97 DO - https://doi.org/10.2991/amitp-16.2016.97 ID - Zhang2016/09 ER -