Low power algorithm for shoe mounted inertial navigation system

Abstract

Inertial navigation systems are adopted in various scenarios where systems cannot rely on external reference source. In this work to create an algorithm for human movement tracking acceleration and magnetic field signals are used. The present work addresses problems of human walking pattern recognition, heading estimation, specially with possible tilt compensation and step length estimation from accelerometer data. All problem solutions are combined into one algorithm. Finally algorithm was tested to check the performance of low power algorithm.