周于涛, 吴华意, 成洪权, 郑杰, 李学锡. 结合自注意力机制和结伴行为特征的行人轨迹预测模型[J]. 武汉大学学报 ( 信息科学版), 2020, 45(12): 1989-1996. DOI: 10.13203/j.whugis20200159
引用本文: 周于涛, 吴华意, 成洪权, 郑杰, 李学锡. 结合自注意力机制和结伴行为特征的行人轨迹预测模型[J]. 武汉大学学报 ( 信息科学版), 2020, 45(12): 1989-1996. DOI: 10.13203/j.whugis20200159
ZHOU Yutao, WU Huayi, CHENG Hongquan, ZHENG Jie, LI Xuexi. Pedestrian Trajectory Prediction Model Based on Self-Attention Mechanism and Group Behavior Characteristics[J]. Geomatics and Information Science of Wuhan University, 2020, 45(12): 1989-1996. DOI: 10.13203/j.whugis20200159
Citation: ZHOU Yutao, WU Huayi, CHENG Hongquan, ZHENG Jie, LI Xuexi. Pedestrian Trajectory Prediction Model Based on Self-Attention Mechanism and Group Behavior Characteristics[J]. Geomatics and Information Science of Wuhan University, 2020, 45(12): 1989-1996. DOI: 10.13203/j.whugis20200159

结合自注意力机制和结伴行为特征的行人轨迹预测模型

Pedestrian Trajectory Prediction Model Based on Self-Attention Mechanism and Group Behavior Characteristics

  • 摘要: 理解并准确预测行人的移动轨迹,对提高自动驾驶技术的水平,减少交通事故的发生有重要的意义。针对现有轨迹预测方法预测精度不高,对行人交互信息利用不充分等问题,提出了一种结合自注意力机制和结伴行为特征的行人轨迹预测模型,该模型考虑了每个行人的运动信息及其与周围行人的交互作用,使用循环神经网络和图卷积网络分别对行人的行走状态和行人间的交互进行建模。在图卷积网络中,定义图的节点表示行人的运动信息,图的边表示行人之间的交互,使用自注意力机制计算行人间的交互程度。此外,为了增加模型捕捉结伴行走行为特征的能力,提高对该类轨迹预测的精度,提出了同伴损失函数的概念。在公共数据集上的实验表明,该模型在预测精度上相比其他方法有较大的提升。

     

    Abstract: Understanding and accurately predicting the trajectory of pedestrians is of great significance to improve the level of auto-driving technology and reduce the occurrence of traffic accidents. Aiming at the problems of low prediction accuracy and insufficient utilization of pedestrian interaction information, a pedestrian trajectory prediction model based on self-attention mechanism and group behavior characteristics is presented. The model considers the movement information of each pedestrian and its interaction with the surrounding pedestrians. The recurrent neural network and graph convolutional network are used to model the pedestrian's walking state and the pedestrian's interaction separately. In the graph convolutional network, nodes of the graph represent the movement information of pedestrians, and edges of the graph represent the interaction between pedestrians. Self-attention mechanism is used to calculate pedestrian interaction. In addition, in order to increase the ability of the network to capture the group walk behavior and decrease the prediction error, a peer loss function is proposed. Experiments on public datasets show that the proposed model in this paper improves prediction accuracy significantly.

     

/

返回文章
返回