中国安全科学学报 ›› 2026, Vol. 36 ›› Issue (8): 225-232.doi: 10.16265/j.cnki.issn1003-3033.2026.08.0054

• 公共安全与应急管理 • 上一篇    下一篇

面向工业仓储安全的风险感知概率轨迹预测方法

李鑫1(), 宋轩宇1, 王乐瑶2, 未宗帅1, 柯巍3, 佟瑞鹏1,**()   

  1. 1 中国矿业大学(北京) 应急管理与安全工程学院, 北京 100083
    2 首都经济贸易大学管理工程学院, 北京 100070
    3 湖北省烟草公司十堰市公司, 湖北 十堰 442099
  • 收稿日期:2026-04-11 修回日期:2026-06-14 出版日期:2026-08-28
  • 通信作者:
    **佟瑞鹏(1977—),男,黑龙江穆棱人,博士,教授,主要从事行为安全管理、职业心理健康、环境风险评估等方面的研究。E-mail:
  • 作者简介:

    李鑫 (1997—),男,山西翼城人,博士研究生,主要研究方向为智能工业安全、风险识别、安全管理等。E-mail:

    柯巍 工程师

  • 基金资助:
    国家自然科学基金资助(52074302); 中国烟草总公司湖北省公司科技项目(2024SY3RCJSWL2C031)

Risk-aware probabilistic trajectory prediction method for industrial warehouse safety

Li Xin1(), Song Xuanyu1, Wang Leyao2, Wei Zongshuai1, Ke Wei3, Tong Ruipeng1,**()   

  1. 1 School of Emergency Management and Safety Engineering, China University of Mining and Technology-Beijing, Beijing 100083, China
    2 School of Management and Engineering, Capital University of Economics and Business, Beijing 100070, China
    3 Shiyan Company of Hubei Tobacco Company, Shiyan Hubei 442099, China
  • Received:2026-04-11 Revised:2026-06-14 Published:2026-08-28

摘要:

为提高工业仓储安全(IWS)关键场景中未来轨迹预测、碰撞风险识别及概率评估的可靠性,提出风险感知概率轨迹预测(RAPTP)方法。该方法利用二维特殊欧氏群(SE(2))等变状态空间编码器表征多智能体交互,结合流匹配与神经危险场联合生成未来轨迹并估计碰撞风险,引入分布鲁棒优化和蒙德里安共形预测进行概率校准。同时,构建IWS基准数据集,并基于nuScenes公开数据集、IWS基准数据集和独立采集的湖北某物流中心数据集开展对比、消融及零样本迁移实验。结果表明:RAPTP在nuScenes数据集上的最小最终位移误差(minFDE)为1.50 m,较EqMotion降低16.2%;在IWS基准上,未遂事故检测曲线下面积(AUC)较碰撞时间(TTC)基线提高34.4%,预警提前时间由1.5 s提升至3.8 s,90%目标覆盖率下预测区间覆盖率达94.2%;在湖北数据集遮挡场景中的检测AUC达89.2%。RAPTP能够兼顾轨迹预测、碰撞预警、概率校准和跨场景迁移性能。

关键词: 工业仓储安全(IWS), 风险感知概率轨迹预测(RAPTP), 流匹配, 神经危险场, 碰撞风险

Abstract:

To improve trajectory prediction accuracy, collision risk identification, and probabilistic assessment reliability in industrial warehouse scenarios, a RAPTP method is proposed. RAPTP employs an SE(2)-equivariant state-space encoder to model multi-agent interactions, combines flow matching with a neural hazard field to jointly generate future trajectories and estimate collision risk, and integrates distributionally robust optimization (DRO)with Mondrian conformal prediction for probability calibration. An IWS benchmark is also constructed, followed by comparative, ablation, and zero-shot transfer experiments. RAPTP achieves a minimum Final Displacement Error(minFDE) of 1.50 m on nuScenes, 16.2% lower than EqMotion. On the IWS benchmark, the near-miss detection Area Under the receiver operating characteristic Curve(AUC) is 34.4% higher than Time-to-Collision(TTC) baseline, the early warning time increases from 1.5 s to 3.8 s, and the prediction interval coverage reaches 94.2% at a target coverage level of 90%. In occlusion scenarios of the Hubei dataset, the detection AUC reaches 89.2%. These results demonstrate that RAPTP achieves effective trajectory prediction, collision warning, probability calibration, and cross-scenario transferability.

Key words: industrial warehouse safety (IWS), risk-aware probabilistic trajectory prediction(RAPTP), flow matching, neural hazard field, collision risk

中图分类号: