Low-density plantar-pressure sensing for gait-event and gait-phase detection: a scoping review of sensor topology, algorithms, validation rigor, robustness, and reproducibility

Abstract

Preprint — not peer reviewed. Submitted to Gait & Posture on 29 August 2026. This scoping review maps 225 included reports on low-density plantar-pressure/contact sensing (≤16 discrete sensing elements per foot) for human walking gait-event and gait-phase detection. It characterizes sensor density and anatomical topology, detector algorithms, reference-standard/comparator validation, real-time implementation, pathological-gait evaluation, robustness, and reproducibility. The initial review corpus comprised 2,186 screened records. Of 397 records advanced to full-text retrieval, 294 identity-matched complete full texts were assessed, 103 were not retrieved or remained unresolved after bounded rescue, 70 assessed full texts were excluded, and 224 reports were included. A final update through 29 August 2026 identified one newly eligible report that was independently assessed and included by both human reviewers, yielding a final synthesis of 225 reports. Initial pre-consensus full-text agreement was 96.6% (Cohen's κ = 0.90625). Overall, 59.6% of included reports used 1–4 sensors per foot and 82.2% used 1–8. Real-time or online implementation was reported in 78.2%, pathological gait evaluation in 68.4%, reference-standard/comparator validation in 70.7%, data availability in 17.3%, and code availability in 2.7%. Very sparse configurations were therefore common within the included low-density evidence base. However, this review does not establish the prevalence or superiority of low-density sensing relative to dense plantar-pressure arrays. Supplementary material, the curated 225-report study table, and aggregate source data are included with this Zenodo record.