Real-Time Gait Phase Estimation Based on Textile Integrated Ferroelectrets and Adaptive Oscillators

Accurate real-time gait phase estimation is essential for the effective control of assistive devices such as exoskeletons and prostheses. Traditional methods, including force plates and inertial measurement units (IMUs), suffer from various limitations such as calibration requirements, drift, and dependency on laboratory settings. In this work, we propose a novel approach that integrates ferroelectret sensors into textile garments, leveraging their high sensitivity and dynamic force measurement capabilities. The sensor signals are processed using an adaptive oscillator (AO)-based algorithm to generate a continuous gait phase estimate. The system is evaluated in a treadmill experiment where a participant walks at varying speeds (4-6.5 km/h). Results demonstrate that the estimated gait phase successfully phase-locks after approximately four strides, maintaining synchronization across all tested speeds with a mean absolute phase error of 0.23 rad. The adaptive thresholding method reliably detects gait events, while the AO structure ensures smooth phase estimation. Although sensor placement and muscle crosstalk introduce complexities compared to traditional approaches, the unobtrusive textile integration offers significant advantages for user comfort and mobility. Future work will focus on multi-sensor fusion and improved signal processing to enhance robustness and accuracy. This study presents a proof of concept for textile-integrated ferroelectret sensors as a promising alternative for real-time gait phase detection in assistive technologies.