基于乘客主观感知的公交客流拥塞量化表征模型

Subjective Perception Based Passenger Congestion Quantification for Bus Operation

为了综合量化表征公交走廊客流承载设施及工具的空间拥挤性和客流运送时间效率,并科学评价客观技术指标和主观乘客感知对公交系统运行状态的敏感度,提出客流拥塞的概念,并用客流拥塞指数量化表征公交系统运行状态。通过解析客流拥塞与出行时效的关联,并分析公交出行阶段特征,基于"时效延长"思想构建客流拥塞量化表征模型,用于系统量化乘客对出行时间和空间拥挤的主观感知;选取影响出行时效的核心要素作为特征变量(车厢客流密度、站台乘客密度、区间乘车时间和站点候车时间),采用直观类比打分SP调查法获取数据用于估计模型参数;基于北京西三环公交走廊实测数据,分析全天区间客流拥塞指数时空分布特征和走廊客流拥塞指数动态演化趋势。研究结果表明:北京西三环公交走廊客流拥塞高峰阈值为0.193 9,客流拥塞状态存在明显的方向特征,下行方向客流拥塞高峰状态更显著且持续时间更长;4个特征变量灵敏度系数均值分别为0.449 2,0.165 2,1.427 1和0.408 3,即区间乘车时间为客流拥塞指数最显著的影响因子,而站台乘客密度的影响程度最小,模型识别公交客流拥塞成因的能力得以体现;该模型能够综合全面地反映公交走廊客流拥塞时空分布和动态演化趋势,并能够应用于公交运行状态改善措施研究。

To comprehensively quantify crowding and passenger transport efficiency of the equipment and facilities of a bus corridor,and scientifically evaluate the sensitivity of objective technical index and subjective passenger perception to the level of the bus running condition,the conception of passenger congestion and passenger congestion index models was proposed.From the analysis of the correlation between subjective passenger travel time and passenger congestion and the characteristics of bus travel,the passenger congestion quantification model was formulated with a time multiplier for systematically evaluating subjective perception,travel time,and crowding.A visual analogy scoring SP survey was designed for model parameter estimation data collection,incorporating four characteristic variables（passenger density in carriage,passenger density on platform,travel time,and waiting time）affecting perception time.Applying the practical data of a Beijing Xisanhuan bus corridor,the all-day temporal-spatial distribution of section passenger congestion index and the dynamic evolutionary trend of the corridor passenger congestion index were studied.According to the results,based on the peakstate threshold（the value of which is 0.193 9）,the passenger congestion condition of the Xisanhuan bus corridor presents significantly directional distinction,and the peak-state is more remarkable and lasts longer in the downward direction.Furthermore,the average sensitivity coefficients of the four characteristic variables are 0.449 2,0.165 2,1.427 1,and 0.408 3,respectively,and travel time is the most significant factor of the passenger congestion index while passenger density on platform is the least,showing that the model is capable of distinguishing causes of formation.This model has the advantage of comprehensively analyzing the temporalspatial distribution and dynamic evolutionary trend of passenger congestion in a bus corridor,and may hopefully be applied to bus running condition improvement measures research.