The development of a smart-carpet for plantar pressure analysis using optical distributed sensing and machine learning

Abstract

Plantar pressure analysis is an important tool in the scope of healthcare as it offers valuable insights into foot structure, corporal balance and overall health. In this context, this paper presents a smart-carpet embedded with an optical fiber acting as a distributed sensing element for plantar pressure analysis and shape reconstruction. Optical Frequency Domain Reflectometry (OFDR) was employed, and the measured signals were cross-correlated with the unloaded response to obtain the spectral shift over the fiber length. An image processing algorithm that utilizes edge detection and clustering was implement to estimate the pressure profile across the footprint. After signal acquisition and processing, a spatial resolution of $5\mathrm{mm}$ was achieved. The right footprints of four healthy subjects were analyzed using the proposed technology, and the spectral shift quality remained within a reliable range for subjects weighing up to 84 kg. To map the spectral shift to pressure, a characterization procedure was performed, resulting in a linear fit with an $R^2$ value of 0.92 and a sensitivity of 1.83 Pa/GHz. The average pressure distributions across the hindfoot, midfoot and forefoot were 42.7%, 24.38% and 32.80%, respectively, and footprint lengths were estimated with relative errors ranging from 2.98% to 5.38%. The proposed solution contributes towards advancements in plantar pressure analysis and development of personalized 3D-printed insoles using distributed optical sensing.