LSTM based trajectory prediction model for cyclist utilizing multiple interactions with environment

作者：

Highlights：

• A unified LSTM network framework models cyclist's interaction with the environment.

• Road key points address the interaction with road and static obstacles.

• The focal attention mechanism improves LSTM by focusing on more relevant features.

• MI-LSTM acquires the knowledge of interactions and outperforms the typical state-of-the-art approaches in most cases.

摘要

•A unified LSTM network framework models cyclist's interaction with the environment.•Road key points address the interaction with road and static obstacles.•The focal attention mechanism improves LSTM by focusing on more relevant features.•MI-LSTM acquires the knowledge of interactions and outperforms the typical state-of-the-art approaches in most cases.

论文关键词：Trajectory prediction,Interaction,Cyclist,LSTM,Focal attention mechanism

论文评审过程：Received 7 May 2020, Revised 6 October 2020, Accepted 7 December 2020, Available online 31 December 2020, Version of Record 31 December 2020.

论文官网地址：https://doi.org/10.1016/j.patcog.2020.107800