电子学报 ›› 2022, Vol. 50 ›› Issue (8): 1905-1916.DOI: 10.12263/DZXB.20211613

• 学术论文 • 上一篇    下一篇

基于全局自适应有向图的行人轨迹预测

孔玮, 刘云, 李辉(), 崔雪红, 杨浩冉   

  1. 青岛科技大学信息科学技术学院,山东 青岛 266061
  • 收稿日期:2021-12-03 修回日期:2022-02-07 出版日期:2022-08-25
    • 通讯作者:
    • 李辉
    • 作者简介:
    • 孔 玮 女,1986年生,山东济南人.青岛科技大学信息科学技术学院博士研究生.主要研究方向为计算机视觉、轨迹预测.
      刘 云 男,1962年生,山西太原人.青岛科技大学信息科学技术学院教授、博士生导师.主要研究方向为计算机视觉、轨迹预测等.
      李 辉 男,1984年生,河南平顶山人.青岛科技大学信息科学技术学院副教授、硕士生导师.主要研究方向为计算机视觉、多目标跟踪与轨迹预测等.
      崔雪红 女,1978年生,山东菏泽人.青岛科技大学信息科学技术学院高级实验师.主要研究方向为计算机视觉、多目标检测与跟踪.
      杨浩冉 女,1997年生,河北邯郸人.青岛科技大学信息科学技术学院硕士研究生.主要研究方向为3D点云目标跟踪.
    • 基金资助:
    • 国家自然科学基金(61702295);山东省高等学校优秀青年创新团队计划(2019KJN047)

Pedestrian Trajectory Prediction Based on Global Adaptive Directed Graph

KONG Wei, LIU Yun, LI Hui(), CUI Xue-hong, YANG Hao-ran   

  1. School of Information Science and Technology, Qingdao University of Science and Technology, Qingdao, Shandong 266061, China
  • Received:2021-12-03 Revised:2022-02-07 Online:2022-08-25 Published:2022-09-08
    • Corresponding author:
    • LI Hui
    • Supported by:
    • National Natural Science Foundation of China(61702295);Outstanding Youth Innovation Team Project for Universities of Shandong Province(2019KJN047)

摘要:

由于行人交互的复杂性和周围环境的多变性,行人轨迹预测仍是一项具有挑战性的任务.然而,基于图结构的方法建模行人之间的交互时,存在着网络感受野小、成对行人间的相互交互对称、固定的图结构不能适应场景变化的问题,导致预测轨迹与真实轨迹偏差较大.为了解决这些问题,本文提出一种基于全局自适应有向图的行人轨迹预测方法(pedestrian trajectory prediction method based on Global Adaptive Directed Graph,GADG).设计全局特征更新(Global Feature Updating,GFU)和全局特征选择(Global Feature Selection,GFS)分别提升空间域和时间域的网络感受范围,以获取全局交互特征.构建有向特征图,定义行人间的不对称交互,提高网络建模的方向性.建立自适应图模型,灵活调整行人间的交互关系,减少冗余连接,增强图模型的自适应能力.在ETH和UCY数据集上的实验结果表明,与最优值相比,平均位移误差降低14%,最终位移误差降低3%.

关键词: 轨迹预测, 自适应图, 有向图, 感受野, 行人轨迹, 图卷积

Abstract:

Due to the complexity of pedestrian interaction and the variability of the surrounding environment, pedestrian trajectory prediction is still a challenging task. However, when modeling pedestrian interaction based on graph structure, there are some problems, such as small sensing field of the network, symmetrical interaction between pedestrians, and fixed graph structure that can not adapt to scene changes, which lead to a large deviation of the predicted trajectory from the real trajectory. To solve these problems, a pedestrian trajectory prediction method based on global adaptive directed graph is proposed. Global feature updating(GFU) and global feature selection(GFS) are designed to improve the perception range in spatial and temporal domain respectively and get global interaction features. A directed feature graph is constructed to define the asymmetric interaction between pedestrians and improve the directionality of network modeling. An adaptive graph model is established to flexibly adjust the relationship between pedestrians, reduce redundant connections and enhance the adaptive ability of the graph. The experimental results on ETH and UCY datasets show that comparing with the optimal value, the average displacement error is reduced by 14% and the final displacement error is reduced by 3%.

Key words: trajectory prediction, adaptive graph, directed graph, sensing field, pedestrian trajectory, graph convolution

中图分类号: