Abstract | ||
---|---|---|
Estimating the short-term online taxi travel time is an important content in urban planning and navigation forecasting systems. When estimating the taxi travel time, we need to take many factors, such as temporal correlation, spatial dependency, and external factors, into consideration. In this paper, we propose a model named DeepSTTE (Short-term Travel Time Estimation) to estimate the short-term online taxi travel time. Firstly, the model integrates external factors using the embedding method. Further, we leverage the classical convolution networks to obtain the spatial feature information of the original GPS trajectory, and use the temporal convolutional networks (TCN) to obtain the temporal characteristics. Finally, we estimate the online taxi travel time of the entire path by the auxiliary learning part. We perform lots of experiments with real datasets, showing that our model DeepSTTE reduces the errors and performs better than the current methods in estimating the travel time. |
Year | DOI | Venue |
---|---|---|
2020 | 10.1007/978-3-030-59019-2_3 | WASA |
DocType | Citations | PageRank |
Conference | 1 | 0.36 |
References | Authors | |
0 | 5 |