A Step Length Estimation based on Motion Recognition and Adaptive Gait Cognition using as Smartphone

Abstract

The step length estimation plays a key role in measuring the user's location on a PDR system. Although its importance becomes greater with a wide spread of smart phone, improving the accuracy of the step length estimation has been left as a task to be resolved yet. It is because many factors of an individual human movement on specific motion should be taken into account. Thus, various motions of human need to be figured out to improve the accuracy of step length estimation. In this paper, we propose a motion based step length estimation algorithm using sensory functions of a smart phone. In the proposed algorithm, the motions of user are identified using hybrid model of decision tree (DT), artificial neural network (ANN) and support vector machine (SVM). Map-based in-flight calibration is also employed to analyze the gait characteristic of user and enhance the step length estimation without any RF assist. In order to verify the performance of the proposed algorithm, we performed experiments on 5 subjects in indoor and demonstrated its improved accuracy on estimating the step length.