Evaluating the Effectiveness of Computer Vision Systems Mounted on Shared Electric Kick Scooters to Reduce Sidewalk Riding

Oshanreh, Mohammad Mehdi; MacKenzie, Donald; Malarkey, Daniel · 2023 · ROSA P / Connected Cities for Smart Mobility toward Accessible and Resilient Transportation Center (C2SMART)

archive: archived pipeline: cataloged verified

Get this paper ↗ (full text — opens at the source; we link to it, we don't host it)

Summary

This study evaluates the effectiveness of computer vision systems mounted on shared electric kick scooters (e-scooters) in reducing sidewalk riding, a behavior that creates safety conflicts with pedestrians and violates local ordinances in many cities. The research addresses the challenge of enforcing riding regulations without dedicated infrastructure, testing whether real-time feedback and speed limitations can alter rider behavior. The study was conducted in Santa Monica, California, where sidewalk riding is prohibited, utilizing data from Spin, a micromobility provider. The researchers conducted a quasi-experiment from November 23, 2022, to February 14, 2023, involving 100 e-scooters equipped with AI cameras capable of detecting surface types (sidewalk, street, or bike lane). Fifty scooters were assigned to a "feedback" group, which received auditory alerts, in-app notifications, and speed reductions when riding on sidewalks. The other 50 scooters constituted a "no-feedback" control group, where these mechanisms were disabled, though surface detection continued. The study analyzed 488 trips, calculating time and distance metrics based on GPS coordinates and event logs. Additionally, riders completed surveys to assess their perceptions of the feedback and reasons for sidewalk riding. Statistical analysis using Empirical Cumulative Distribution Function (ECDF) plots and Kolmogorov-Smirnov tests revealed that the feedback mechanisms significantly reduced sidewalk riding. Riders on feedback-enabled scooters spent 22% less time and 26% less distance on sidewalks compared to the control group, while spending 5% more time on streets. The feedback also significantly reduced the length and duration of individual sidewalk riding segments. Binary logistic regression models further demonstrated that feedback decreased the likelihood of riders transitioning onto sidewalks when previously riding on streets or bike lanes. Survey results indicated that riders were aware of the feedback and that it influenced their future likelihood of using the service. The findings suggest that onboard computer vision systems combined with auditory feedback and speed limitations are effective tools for guiding e-scooter riders toward legal riding surfaces. By reducing sidewalk usage, these interventions can mitigate conflicts between pedestrians and micromobility users, thereby enhancing safety and compliance with city regulations. The study supports the integration of behavioral strategies, such as real-time driver feedback, into vehicle design to address urban mobility challenges where infrastructure solutions are insufficient.

Key finding

Real-time auditory feedback and speed limitations on AI-equipped e-scooters significantly reduced the proportion of trip time and distance spent on sidewalks compared to scooters without such feedback.

Methodology

naturalistic

Sample size: 488

Provenance

The full processing record for this entry. Every stage of this paper's journey through the pipeline is logged — what ran, with which tool and model, how many attempts it took, and when it last completed. Discovered via bulk_ingest_rosap on 2026-05-23 (6 acquisition events logged).

Summary generated by qwen3.6-27b-prismaquant on 2026-06-10; verification: verified.

Topics

Ranked by relevance to this paper. Hover a topic for its definition.

• micromobility e scooter