摘要
在轨道交通客流实时监视业务中,车站实时客流告警的目的是帮助业务人员快速、准确地定位到网络中可能存在大客流风险的车站,达到提示预警的作用。如果告警阈值偏小,则告警频率会增加,使业务人员无法准确判断最需要关注的大客流风险车站,并对告警提示产生麻木心态;如果报警阈值偏大,则车站的突发大客流风险有可能未被监测到,从而无法及时采取应对措施。因此,需要提出一套科学、合理、适用性强的告警阈值。研究利用传统统计学方法和基于孤立森林的异常检测方法训练设定告警阈值,并通过应用效果测试验证对两种算法进行比选,确认利用孤立森林异常值判别思路训练的告警阈值更满足业务目标。研究成果按照“一站一方案”“一个时段一方案”的原则,为北京市轨道交通指挥中心路网调度指挥平台提供车站实时进站量与出站量告警阈值,支撑客流实时监视与大客流风险预警。
In the rail transit real-time passenger flow monitoring business,a real-time station passenger flow alarm assists operators in quickly and accurately locating stations with potential significant passenger flow risks,enabling early warning.If the alarm threshold is too low,then the alarm frequency increases,making it difficult for the operator to accurately identify high-risk stations requiring attention and potentially leading to alarm fatigue.Conversely,if the alarm threshold is too high,then the risk of sudden large passenger flow at the station may go unnoticed,and timely responses may not be possible.Therefore,the primary objective of this study was to propose a set of scientific,reasonable,and applicable alarm thresholds.This study employed traditional statistical methods and isolation forest methods based on machine learning to train and set the alarm threshold.Furthermore,an application-effect test was used to compare the two algorithms.Training the outliers in the isolation forest ensures better alignment of the alarm threshold with business goals.Adhering to the principle of“one station,one plan”and“one time period,one plan”,this study provides the alarm threshold for real-time inbound and outbound volumes at stations for the Beijing Metro Network Control Center(BMNCC)dispatching emergency command,passenger guidance,and information service platform.This supports real-time passenger flow monitoring and risk early warning for large passenger flow.
作者
王月玥
孙琦
钟厚岳
WANG Yueyue;SUN Qi;ZHONG Houyue(Beijing Metro Network Control Center,Beijing 100101)
出处
《都市快轨交通》
北大核心
2023年第3期71-76,共6页
Urban Rapid Rail Transit
基金
北京市基础设施投资有限公司2020年度科研项目(2020-ZH-04)。
关键词
轨道交通
孤立森林
告警阈值
客流监视
rail transit
Isolation Forest
alarm threshold
passenger flow monitoring
