期刊
ROBOTICS AND AUTONOMOUS SYSTEMS
卷 161, 期 -, 页码 -出版社
ELSEVIER
DOI: 10.1016/j.robot.2022.104352
关键词
Human trajectory prediction; Social interactive feature; STGCNN; Attention mechanism; Speculation
A Spatio-Temporal Graph Convolution Neural Network based Social Interaction Model (STGCNN-SIM) is proposed to accurately predict human trajectories by utilizing historical and speculated trajectory information to extract social interactive features and model interaction behaviors. Three social interactive features, including relative distance, angle between velocity vectors, and angles between velocity vectors and distance vector, are explicitly extracted from observed and speculated trajectories. STGCNN-SIM utilizes these features to model interactions with surroundings and employs an attention mechanism to improve the model's performance. Experimental results on three public datasets demonstrate that STGCNN-SIM achieves higher accuracy and stability compared to state-of-the-art methods.
Accurate human trajectory prediction is still challenging due to the complicated interactions with surroundings. A Spatio-Temporal Graph Convolution Neural Network based Social Interaction Model (STGCNN-SIM) is proposed to address this challenge. In addition to historical trajectory information, the presented method employs the speculated trajectories in the future to extract social interactive features and model interaction behaviors. Three social interactive features are extracted explicitly from the observed and speculated trajectories: (1) the relative distance, (2) the angle between the velocity vectors of two interacting partners, and (3) the angles between the velocity vectors of interacting partners and the distance vector. STGCNN-SIM utilizes these social interactive features to model interactions with surroundings in the historical and speculated stages. Then an attention mechanism is adopted to improve the model by focusing on more relevant features. Experimental results on three public datasets demonstrate that STGCNN-SIM achieves higher accuracy and stability than the state-of-the-art methods.(c) 2022 Elsevier B.V. All rights reserved.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据