Smartphone based gait analysis using STFT and wavelet transform for indoor navigation

摘要: In this paper, we propose a frequency domain analysis for characterizing the walking gait in context of indoor navigation, without assuming that sensors are rigidly attached to body. Firstly, is performed using Short Time Fourier Transform (STFT) since statistical properties signal changing over time but assumed contant hort window. Globally STFT can extract step/stride frequency, found non optimal fast motion transitions. Wavelet (WT) then introduced. Contrary STFT, WT uses size-adjustable window, which offers more advantages human features extraction. When period interest comprises high window short, while when local area low size enlarged. This propriety be critical our smartphone based analysis. Experimental assessment with Nokia Lumia 920 and foot mounted MEMS grade inertial used as reference. These results encouraging designing robust adaptable real-time detection solution navigation.