Kinematic Characterization of Risky Riding Behavior of On-Demand Food-Delivery Motorcyclists in Taiwan

The rising on-demand food-delivery (ODFD) industry has raised road safety concerns because of the often aggressive and risky driving behavior among ODFD drivers, especially those on motorcycles. This study aims to characterize their aggressive and risky driving behavior with respect to microscopic kinematic characteristics and vehicular interactions. We employed traffic videos collected by unmanned aerial vehicles at two intersections in Taipei, Taiwan, with a 0.1 s time resolution. We extracted vehicular trajectory data using artificial intelligence-based video recognition algorithms to obtain microscopic kinematic variables. We compared microscopic traffic flow characteristics (i.e., speed, lateral velocity, and acceleration) and microscopic interactions (i.e., weaving maneuver frequency, safety gap, and time-to-collision) presented by ODFD and non-ODFD drivers. In addition, we compared the difference in their driving behaviors between non-meal-peak hours and meal-peak hours, hypothetically caused by the platform-employed incentive program. We found that, compared to non-ODFD drivers, ODFD drivers could be more likely to perform aggressive and risky driving, indicated by their faster longitudinal and lateral speed, harsher acceleration/deceleration, more frequent weaving maneuvers, shorter safety gaps, and shorter time-to-collision. Also, their aggressive and risky driving behavior could more likely occur during meal-peak hours. To our knowledge, this research is the first study using naturalistic traffic data to investigate the revealed driving behavior of ODFD drivers, contributing to kinematic/quantitative understandings of ODFD drivers’ aggressive and risky driving. Based on these findings, policymakers and platform companies can prescribe countermeasures and devise training programs to improve public road safety and the occupational safety of ODFD drivers, a vulnerable occupational group emerging worldwide.