stride_template {adeptdata}

Walking Stride Pattern Templates

Description

Walking stride pattern templates derived from raw accelerometry data collected at four body locations: left wrist, left hip, left ankle, and right ankle.

Usage

stride_template

Format

A list with four named elements:

• left_wrist,

• left_hip,

• left_ankle,

• right_ankle.

Each of the above is a five-element list of matrix objects. The matrices are collection of (sub)population-specific stride pattern templates. For example,

• stride_template$left_wrist[[1]] is a 1 x 200 matrix of one population-specific stride template derived from accelerometry data collected at left wrist.

• stride_template$left_wrist[[2]] is a 2 x 200 matrix of two distinct subpopulation-specific stride templates derived from accelerometry data collected at left wrist. Each row is a one subpopulation-specific stride template.

• stride_template$right_ankle[[5]] is a 5 x 200 matrix of five distinct subpopulation-specific stride templates derived from accelerometry data collected at right ankle.

Details

Raw accelerometry data used to derive walking stride pattern templates were collected as a part of the study on Identification of Walking, Stair Climbing, and Driving Using Wearable Accelerometers, sponsored by the Indiana University CTSI grant and conducted at the Department of Biostatistics, Fairbanks School of Public Health at Indiana University. The study was led by Dr. Jaroslaw Harezlak, assisted by Drs. William Fadel and Jacek Urbanek. Study enrolled 32 healthy participants between 23 and 52 years of age. Participants were asked, among others, to perform self-paced, undisturbed, outdoor walking on the sidewalk. Accelerometry data were collected at four body locations: left wrist, left hip, left ankle, and right ankle.

To derive empirical stride pattern, firstly, from each body location, 642 data segments corresponding to individual walking strides were manually segmented. Secondly, Vector Magnitude (VM), which is a univariate summary of three-dimensional time-series of raw accelerometry data, was computed. Lastly, 642 univariate vectors of VM were interpolated to have the same vector length, scaled, and clustered into 1,2,3,4 and 5 clusters via correlation clustering. The vectors obtained as point-wise means within each cluster were defined to be subpopulation-specific stride pattern templates, respectively.