Author(s): YU, T., FUSCO, A., GUIDOTTI, F., FUCHS, P.X., Institution: NATIONAL TAIWAN NORMAL UNIVERSITY, Country: TAIWAN, Abstract-ID: 738

INTRODUCTION:
Cyclists are vulnerable in mixed traffic, and rapid growth of e-bikes and connected mobility introduces new safety challenges. Cycling-safety technologies (bike-integrated sensors, cyclist-worn equipment, cyclist-facing software/apps) are developing rapidly, but evidence is scattered across engineering and health literature, uses heterogeneous outcomes, and often lacks comparable reporting across settings. A scoping review is therefore needed to map which technologies have been evaluated, which outcomes and methods have been used, and where evidence gaps remain. We aimed to synthesize and structure the current evidence base to inform future research priorities and bicycle-centered safety innovation.
METHODS:
We conducted a scoping review following the PRISMA Extension for Scoping Reviews [1] and the JBI Manual for Evidence Synthesis [2]. Eligibility was defined using the Population–Concept–Context framework: cyclists (all bicycle types), bicycle- or cyclist-centered safety technologies (bike-integrated, cyclist-worn/carried, or cyclist-facing apps), and public-road traffic contexts (urban, peri-urban, or rural; road-relevant laboratory or simulation studies were eligible). We searched Scopus, PubMed, and Embase (2021–2026) with database-adapted syntax. Screening was conducted by two independent reviewers: the primary investigator and a large language model trained on the study’s inclusion and exclusion criteria. Discrepancies were resolved by a third senior reviewer. Data were extracted and inserted into a pre-defined template, following a standardized protocol for evidence mapping.
RESULTS:
After deduplication and exclusion of 41 ineligible records, the final evidence base comprised 211 sources (169 journal articles, 39 conference publications, 3 other sources). Investigation types included 202 original investigations, 7 reviews (e.g., systematic reviews, meta-analyses), and 2 other publications (e.g., commentaries). Among original investigations, the most common settings were traffic or field environments (45%), laboratory or bench testing (33%), and computational simulation (14%). Technology focus was distributed across cyclist-worn equipment (59%), software/apps (32%), and bike-integrated hardware (9%). The three most frequently studied tools were computer vision (23%), IMU (42%), and warning (22%) systems.
CONCLUSION:
This review revealed an evidence base heavily skewed toward cyclist-worn and app-based tools (combined 90.6%), which focused on monitoring in most cases. In contrast, bicycle-integrated hardware (9.4%) remains under-investigated, despite its potential to provide consistent protection independent of user adherence. Consequently, scientific evidence supporting the effectiveness of bicycle-integrated systems remains limited. Future research may prioritize the naturalistic validation of such systems.

REFERENCES:
1. Tricco et al. (2018). Annals of Internal Medicine, 169(7), 467-473.
2. Munn et al. (2023). JBI Evidence Synthesis, 21(3), 592-600.