On Expressive Features for Gait Analysis using Lower Limb Inertial Sensor Data

摘要: Abstract In this paper, we present a method to obtain explicit, expressive and interpretable gait feature signals from an inertial sensor, mounted on any segment of the lower limbs. The proposed is invariant mounting orientation works without magnetometer information, requires no prior knowledge can be used in real-time scenarios. Moreover, constructed are robust for wide variety changing walking speeds directions. We investigate informational content our three lying human sagittal plane with respect phase segmentation problem compare them other commonly signals, such as angular velocity norms accelerations velocities. To end, make use filter-based maximum relevance minimum redundancy algorithm, which classifier-independent selection method. For validating approach, consider data twelve healthy subjects straight curves at self-chosen sensors attached either thigh, shank or foot. Additionally, pressure measuring insoles ground truth toe-off heel-strike events reference. With those transitions, event detection cast into classification problem. support theoretical findings ranking, finally evaluate different choices sets simple linear vector machine classifier online fashion superior results signals.