Skip to main content
QUICK REVIEW

[论文解读] Pedestrians rotation measurement in bidirectional streams

Claudio Feliciani, Katsuhiro Nishinari|arXiv (Cornell University)|Oct 23, 2016
Evacuation and Crowd Dynamics被引用 8
一句话总结

本研究提出一种新方法,通过安装在行人胸部的商用平板电脑中的陀螺仪,直接测量双向人流中行人的身体旋转,实现高精度、实时的偏航运动追踪。结果表明,与单向流相比,双向流(尤其是车道形成期间)导致显著更高的身体旋转;且直接测量的旋转与基于轨迹的估计结果在定性上一致,验证了间接测量方法的可靠性。

ABSTRACT

This study presents an experimental measurement of pedestrians' body rotation in bidirectional streams. A mock-up corridor monitored using a camera placed on azimuthal position is used to study pedestrians' behavior in unidirectional and bidirectional flows. Additionally, a commercial tablet is fixed on the chest of sample pedestrians to examine their body rotation (or yawing) which cannot be obtained using position tracking alone. Angular velocity is recorded and simultaneously stored in a central location using a wireless network, thus allowing the analysis of body movements with a high sampling rate and a limited delay. To investigate the influence of major/minor flow proportion (flow-ratio) on bidirectional streams two different situations were tested: the balanced configuration (with equal flows in both directions) and an unbalanced configuration (with different major and minor flow). Results clearly show that unidirectional flow is more stable compared to the bidirectional case, requiring less time to cross the experimental section and showing a very small amount of rotation during the whole experiment. Both bidirectional configurations showed high values of body rotation, in particular during lane formation and dissolution. Finally, rotation directly measured on pedestrians' body was compared with the one obtained indirectly by analyzing pedestrians' trajectories. The comparison shows that, at least from a qualitative point of view, both methods are in agreement, thus suggesting that even properties which can only be measured by motion sensing could be obtained indirectly through the analysis of trajectories. Concluding, it has been suggested that while lanes help smooth out bidirectional flows, larger instabilities are observed compared to the unidirectional case. Lane separation and/or appropriate guidance are therefore required.

研究动机与目标

  • 直接测量双向人流中行人的身体旋转(偏航运动),克服仅依赖头部位置追踪的局限性。
  • 研究流量比(平衡与非平衡)对身体旋转及流动稳定性的影响。
  • 验证配备嵌入式惯性传感器的商用平板电脑在行人运动科学测量中的适用性。
  • 将直接测量的身体旋转与基于行人轨迹推断的旋转进行比较。
  • 评估身体旋转在双向人流中车道形成与流动不稳定性中的作用。

提出的方法

  • 为50名参与者中的10人配备固定在胸前的商用平板电脑,利用其内置数字陀螺仪记录角速度。
  • 通过无线网络实时流式传输并存储角速度数据,采样率高且延迟极低。
  • 在模拟走廊中开展实验,分别设置单向流(4/0)和两种双向配置(3/1与2/2流量比)。
  • 计算角速度的平均值及随时间的绝对角速度积分,以量化旋转幅度。
  • 通过数学分析比较直接测量的身体旋转与基于行人轨迹推算的旋转。
  • 进行验证测试,确认商用平板电脑在科学运动测量中的准确性和可靠性。

实验结果

研究问题

  • RQ1在受控走廊环境中,单向流与双向流中行人的身体旋转有何差异?
  • RQ2流量比失衡(如3/1与2/2)在多大程度上影响行人身体旋转的幅度?
  • RQ3能否仅通过行人轨迹数据可靠地估计直接由陀螺仪测得的身体旋转?
  • RQ4身体旋转与双向流中车道的形成和消失过程有何关联?
  • RQ5身体旋转在人流稳定性及自组织结构形成中扮演何种角色?

主要发现

  • 单向流表现出极低的身体旋转(平均 0.363 ± 0.046 rad/s),且更稳定,行人在走廊中穿越速度高于双向条件。
  • 双向流导致显著更高的身体旋转:非平衡3/1配置中为 0.430 ± 0.042 rad/s,平衡2/2配置中为 0.390 ± 0.056 rad/s。
  • 旋转最大值出现在车道形成与消失阶段,表明这些过渡阶段是双向流中的关键不稳定性来源。
  • 绝对角速度积分(旋转范围)在非平衡3/1配置中最高(176.785 ± 23.152 s⁻¹),证实整体运动幅度更大。
  • 直接测量的身体旋转与基于轨迹的估计结果在定性上一致,支持通过间接方法测量旋转等运动特性。
  • 平衡配置中不确定性相对较大,可能源于重复次数较少(3次 vs. 4次),提示需增加实验次数以降低方差。

更好的研究,从现在开始

从阅读论文到最终审阅,大幅缩短您的研究时间。

无需绑定信用卡

本解读由 AI 生成,并经人工编辑审核。