An Adaptive Step Detection Algorithm Based on the State Machine

This paper addresses an adaptive step detection algorithm using inertial sensors commonly found on commodity mobile phone. After analyzing the acceleration data collected by the accelerometer in mobile phone, the step detection is completed by using the state machine combined with other subsidiary conditions. This step detection algorithm can also dynamically change the thresholds based on the acceleration data features to achieve the goal of detecting the steps adaptively. The experimental results show that this step detection algorithm not only has high accuracy, but also solid robustness. It can eliminate the error count caused by the shaking and flipping.