Double-layer ensemble monitoring of non-gaussian processes using modified independent component analysis.
Gait analysis is an important aspect of Biomedical Engineering. In the recent past, researchers have applied several signal processing methods for the analysis of gait activities. Sensors such as accelerometers, gyroscopes and pressure sensors are more commonly used to identify gait activities remotely. Most of the applications have multiple sensors placed on a single board which is used for gait assessment. However, the problem with multiple sensors is the cross talk introduced by one sensor due to another sensor. Some of the applications use a single sensor such as accelerometer with dual axis measuring the gait activity in horizontal and vertical planes. Depending on the orientation of the accelerometer, the two axial outputs could have overlapping spectra which is very difficult to observe. Spectral and temporal filtering is not suitable for this because of overlapping spectra due to simultaneous movements of the foot in the horizontal and vertical planes. To reliably identify the gait activities, there is a need to decompose and separate the two vertical and horizontal acceleration signals. The earlier research has described a novel method which can be used remotely to identify the gait in ITW children. This paper discusses a lab based automated classification method using Blind Source Separation (BSS) technique to identify toe walking gait from normal gait in Idiopathic Toe Walkers (ITW) children. The outcome of the research study reveals that the BSS techniques in association with K-means classifier can suitably distinguish toe-walking gait from normal gait in ITW children with 97.9 ± 0.2% accuracy. © 2014 Elsevier Ltd. All rights reserved.