A temporal fusion transformer for short-term freeway traffic speed multistep prediction

Accurate short-term freeway speed prediction is a key component for intelligent transportation management and can help travelers plan travel routes. However, very few existing studies focus on predicting one-hour ahead or longer freeway speed. In this study, a novel architecture called Temporal Fusion Transformer (TFT) is adopted to predict freeway speed with the prediction horizons from 5 min to 150 min. The TFT can capture short-term and long-term temporal dependence by a multi-head attention mechanism. Moreover, the TFT utilizes the fusion decoder to import various types of inputs which can improve the prediction accuracy. To demonstrate the advantage of the TFT, traffic speed data collected from an interstate freeway in Minnesota are used to train and test the prediction model. The TFT prediction performance is compared with several classic traffic prediction methods, and the results reveal that the TFT performs best in speed prediction when the prediction horizon is longer than 30 min.(c) 2022 Elsevier B.V. All rights reserved.

Automated summary

This study adopts a novel architecture called TFT to predict freeway speed, which can capture short-term and long-term temporal dependence and improve prediction accuracy by incorporating various types of inputs.