Date of Award
Doctor of Philosophy (PhD)
Computational Modeling & Simulation Engineering
Modeling and Simulation
The rapid rise of shared electric scooter (E-Scooter) systems offers many urban areas a new micro-mobility solution. The portable and flexible characteristics have made E-Scooters a competitive mode for short-distance trips. Compared to other modes such as bikes, E-Scooters allow riders to freely ride on different facilities such as streets, sidewalks, and bike lanes. However, sharing lanes with vehicles and other users tends to cause safety issues for riding E-Scooters. Conventional methods are often not applicable for analyzing such safety issues because well-archived historical crash records are not commonly available for emerging E-Scooters.
Perceiving the growth of such a micro-mobility mode, this study aimed to investigate E-Scooter operations and safety by collecting, processing, and mining various unconventional data sources. First, origin-destination (OD) data were collected for E-Scooters to analyze how E-Scooters have been used in urban areas. The key factors that drive users to choose E-Scooters over other options (i.e., shared bikes and taxis) were identified. Concerning user safety tied to the growing usage, we further assessed E-Scooter user guidelines in urban areas in the U.S. Scoring models have been developed for evaluating the adopted guidelines. It was found that the areas with E-Scooter systems have notable disparities in terms of the safety factors considered in the guidelines. Built upon the usage and policy analyses, this study also creatively collected news reports as an alternative data source for E-Scooter safety analysis. Three-year news reports were collected for E-Scooter-involved crashes in the U.S. The identified reports are typical crash events with great media impact. Many detailed variables such as location, time, riders’ information, and crash type were mined. This offers a lens to highlight the macro-level crash issues confronted with E-Scooters. Besides the macro-level safety analysis, we also conducted micro-level analysis of E-Scooter riding risk. An all-in-one mobile sensing system has been developed using the Raspberry Pi platform with multiple sensors including GPS, LiDAR, and motion trackers. Naturalistic riding data such as vibration, speed, and location were collected simultaneously when riding E-Scooters. Such mobile sensing technologies have been shown as an innovative way to help gather valuable data for quantifying riding risk. A demonstration on expanding the mobile sensing technologies was conducted to analyze the impact of wheel size and riding infrastructure on E-Scooter riding experience. The quantitative analysis framework proposed in this study can be further extended for evaluating the quality of road infrastructure, which will be helpful for understanding the readiness of infrastructure for supporting the safe use of micro-mobility systems.
To sum up, this study contributes to the literature in several distinct ways. First, it has developed mode choice models for revealing the use of E-Scooters among other existing competitive modes for connecting urban metro systems. Second, it has systematically assessed existing E-Scooter user guidelines in the U.S. Moreover, it demonstrated the use of surrogate data sources (e.g., news reports) to assist safety studies in cases where there is no available crash data. Last but not least, it developed the mobile sensing system and evaluation framework for enabling naturalistic riding data collection and risk assessment, which helps evaluate riding behavior and infrastructure performance for supporting micro-mobility systems.
"Data-Driven Operational and Safety Analysis of Emerging Shared Electric Scooter Systems"
(2021). Doctor of Philosophy (PhD), Dissertation, Computational Modeling & Simulation Engineering, Old Dominion University, DOI: 10.25777/fgfa-a749