Long-term prediction for high-resolution lane-changing data using temporal convolution network

Abstract

Lane-changing is an important driving behaviour and unreasonable lane changes can potentially result in traffic accidents. Currently, the lane-changing data are often recorded with high resolution, which are not appropriate for some common deep learning approaches. To capture the stochastic time series of high-resolution lane-changing behaviour, this study introduces a temporal convolutional network (TCN) to predict the long-term lane-changing trajectory and behaviour. The lane-changing dataset was collected by the driving simulator at the frequency of 60 Hz. Prediction results show that the TCN can accurately predict the long-term lane-changing trajectory and driving behaviour with shorter computational time compared with two benchmark models including the convolutional neural network (CNN) and long short-term memory neural network (LSTM). The advantages of the TCN are rapid response and accurate long-term prediction, which are important for lane-changing assistance in the advanced driver assistance system.

KEYWORDS:

• Lane changes
• temporal convolution network
• driving behaviour
• trajectory prediction
• time series

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This research was funded by the National Natural Science Foundation of China [Grant Number 71971160], the Shanghai Science and Technology Committee [Grant Number 19210745700] and the Fundamental Research Funds for the Central Universities [grant no. 22120210009].