文章编号： 1672-5328（2026）01-0047-09

张佳珊1，张子豪1，刘晨辉1, 2, 3，谭征宇4
(1. 湖南大学土木工程学院，湖南长沙410082；2. 湖南大学综合交通研究中心，湖南长沙410082；3. 智慧道路 与绿色交通湖南省普通高等学校重点实验室，湖南长沙410082；4. 湖南大学设计艺术学院，湖南长沙410082)

摘要： 外卖骑手交通安全问题日益突出，需系统考察多种因素对其配送特征以及危险驾驶行为的影 响，从而为制定有效的管理措施提供依据。为研究外卖骑手的配送活动特征，利用GPS追踪器和行 驶记录仪分别采集其行驶轨迹和视频数据。首先，基于轨迹数据和视频记录，在划分取餐区和送餐 区的基础上，结合行驶轨迹的起终点位置、平台配送时间要求以及外卖站点服务范围，识别出 3 301 条有效送餐轨迹及其对应的配送活动。随后，计算每次配送活动的总配送时间和送餐轨迹的 超速比例，并利用Beta 随机效应模型分析影响超速行为的关键因素。研究结果表明，外卖骑手每单 平均配送时间为11.2 min，其中等待时间、行驶时间与交付时间分别为2.9 min，4.8 min 和3.5 min。 Beta 随机效应模型分析结果表明，性别、年龄和订单配送距离对超速行为具有显著影响，具体表现 为男性骑手、年龄较大者超速比例更高，且配送距离越远，超速比例也相应提升。

关键词： 交通治理；外卖骑手；自然驾驶数据；配送时间；超速行为；Beta随机效应模型

中图分类号： U491.1

文献标识码：A

Analysis of Delivery Characteristics of Food Delivery Riders Based on Naturalistic Riding Data

Zhang Jiashan1, Zhang Zihao1, Liu Chenhui1, 2, 3, Tan Zhengyu4
(1. College of Civil Engineering, Hunan University, Changsha Hunan 410082, China; 2. Transportation Research Center, Hunan University, Changsha Hunan 410082, China; 3. Hunan Provincial University Key Laboratory of Smart Roads and Green Transportation, Changsha Hunan 410082, China; 4. College of Design, Hunan University, Changsha Hunan 410082, China)

Abstract： The growing prominence of traffic safety issues among food delivery riders necessitates a systematic investigation on the impact of multiple factors on delivery characteristics and risky riding behaviors, thereby providing a basis for developing effective management measures. To examine the delivery activity characteristics of food delivery riders, GPS trackers and dash cameras were used in this study to collect trajectory and video data, respectively. First, based on these data sources and by delineating pick-up and delivery zones, 3,301 valid delivery trajectories and their corresponding delivery activities were identified through an integrated analysis of trajectory origins and destinations, delivery time requirements of platforms, and service area boundaries of food delivery stations. Subsequently, the total delivery time and speeding ratio were calculated for each delivery trajectory, and a Beta random-effects model was applied to analyze the key determinants of speeding behavior. The research results show that the average delivery time per order is 11.2 minutes, including 2.9 minutes of waiting time, 4.8 minutes of riding time, and 3.5 minutes of handover time. The analysis results of the Beta random-effects model further reveals that gender, age, and delivery distance significantly influence speeding behavior. Specifically, male riders and elder riders tend to exhibit higher proportions of speeding, and the proportion of speeding increases with delivery distance.

Keywords： transportation governance; food delivery riders; naturalistic riding data; delivery time; speeding behavior; Beta random effects model