基于改进云模型的地铁站行人拥挤辨识方法

为降低城市地铁站内行人拥挤识别时多指标和多等级带来的不确定性和模糊性,采用改进云模型构建行人拥挤状态辨识模型。基于AHP-熵权法标定指标权重和行人拥挤状态等级,以标定的指标权重和各指标等级阈值作为输入,计算云数字特征值,采用正向正态云发生器,建立模板云和待识别云模型,计算两者之间的隶属度,根据最大隶属度原则辨识车站内各服务设施的行人拥挤状态,最后输出各服务设施的拥挤等级辨识结果。以宁波市鼓楼站为实证对象,对辨识结果进行实证分析。结果表明:采用四级行人拥挤状态划分方法合理可行。

Identification Approach of Pedestrian Crowd in Metro Station Based on Improved Cloud Model

For the purpose of reducing the uncertainty and fuzziness caused by multiple-indicators and multiple-grades in the identification of the pedestrian crowd status in the metro station,an improved cloud model was proposed to identify the pedestrian crowd.The weights of each index and pedestrian crowd grade were calibrated based on the AHP-Entropy method.The calibrated index weights and the thresholds of each grade were used as data input to calculate the cloud digital eigenvalues.A forward normal cloud generator was used to establish the template cloud and the candidate cloud to calculate the degrees of membership between the two.The grade of pedestrian crowd of each service facility in the metro station was calculated according to the principle of the maximum membership degree,and finally the crowd results of each service facility were obtained.The identification approach was verified by a case study in a rail transit station in Ningbo,China.The results indicate that the four-grade division method is reasonable and feasible to reflect pedestrian crowd state.