The Micro-Mobility Sensing Gap: A Systematic Review of Physiological Safety Monitoring from Cycling to E-Scooters
Abstract
The transition from cycling to electric micro-mobility, such as e-scooters, introduces distinct safety risks. While physiological sensing is established for monitoring cyclist exertion, its transferability to high-vibration e-scooter environments remains unclear. This study systematically reviews wearable sensors used to detect stress, fatigue, and exertion in cycling and micro-mobility to identify gaps preventing active safety systems. A PRISMA-guided search of IEEE Xplore, Web of Science, PubMed, Scopus, and ScienceDirect was performed on 2 October 2025 for studies published in 2015-2025. From 273 records, 11 publications representing nine unique studies met the inclusion criteria. Laboratory studies (n=4) utilizing deep learning (CNN-LSTM) achieved high exertion prediction accuracy (F1 86.3-91.7%) but relied on a single redundant dataset (N=27), lacking independent validation. Field studies (n=7) relied on statistical associations between heart rate variability and environmental stress but lacked real-time predictive capabilities. Notably, evidence for automated physiological safety classification in e-scooters is critically underdeveloped. Current models are overfitted to cycling biomechanics and fail to account for e-scooter constraints, such as whole-body vibration. Future research must shift toward Unsupervised Domain Adaptation (UDA) and noise-resilient edge AI architectures to bridge the technological lag in micro-mobility safety.
How strong is the evidence?
Evidence Hub reads this as a systematic review, done in people. Its evidence level is 1 of 5 (Highest · pooled evidence). Pooled evidence from several studies sits at the top of the evidence pyramid. Reported size: 27 participants or records.
Read the paper
Related studies
- Active Learning during Federated Learning for Object Detection
- Fairness-aware federated learning
- Decentralized Federated Learning
- Building a Personally Identifiable Information Recognizer in a Privacy Preserved Manner Using Automated Annotation and Federated Learning
- Preservation of Patient Level Privacy: Federated Classification and Calibration Models
- Efficient explainable federated learning
Go deeper in Evidence Hub
Evidence Hub adds a study data sheet (design, sample size, evidence level, key numbers), an AI summary and a chat grounded in the paper's text.