Design of a Sensor Insole for Gait Analysis

There is an increasing interest in the application of instrumented insoles in sport and medicine to obtain gait information during activities of daily living. Despite the high number of research works dedicated to smart insole design, there is a lack of discussions on strategies to optimize the design of the force sensing electronic acquisition module. Such strategies are needed to achieve a small form factor while maintaining reliable kinetic data acquisition. In the present work, we describe our implementation of a smart insole and demonstrate channel multiplexing to optimize electronic component count. We discuss the details of the analog part, including the analog-to-digital conversion, optimal sampling frequency selection, and methods to reduce errors and influences of component imperfections. We demonstrated a complete framework for insole signal processing developed in Python. We used the insole prototype to collect data from twenty volunteers and implemented a basic algorithm for person recognition. As a result, we achieved a reasonable classification accuracy of 98.75%.