Abstract
Background
Artificial intelligence-enabled sensing and analytical methods increasingly support gait speed measurement beyond conventional laboratory settings, but the development and intellectual structure of this research field have not been comprehensively mapped.
Objective
To map the scientific development of artificial intelligence (AI)-based human gait speed measurement, with emphasis on publication growth, sensing technologies, analytical methods, collaboration patterns, and temporal research trends.
Methods
A descriptive bibliometric analysis was conducted using records indexed in the Web of Science Core Collection, Science Citation Index Expanded, from January 2001 through January 2026. The search retrieved 438 records. Following document-type and language filtering and eligibility screening, 110 publications were included. Bibliometric performance was examined with the bibliometrix R package, and co-authorship, citation, keyword co-occurrence, and temporal overlay maps were generated with VOSviewer.
Results
Publication activity increased markedly after 2020, with 78 publications (70.9%) appearing between 2021 and January 2026. Machine learning, deep learning, inertial measurement units, wearable sensors, gait analysis, and estimation formed the conceptual core of the field. Convolutional neural networks, smartphones, and daily-life monitoring were associated with more recent publications. The co-authorship map showed a connected but concentrated network, while the citation map indicated continuity between earlier and newer studies.
Conclusions
AI-based gait speed measurement is expanding toward automated assessment in real-world settings. Bibliometric growth, however, does not establish measurement accuracy or clinical effectiveness. Future work should prioritize external validation, standardized reporting, device comparability, representative populations, and integration with clinical workflows.