luwangg/accelero2

Human Gait Analysis Using Smartphone Accelerometer

Accelero2

This app compares average gait cycle of both legs and represents it on a scale from 0 to MAX_VALUE.

Smartphone embedded accelerometer is used to take readings voluntary by the user in the app.

Smartphone is kept in the front pocket of jeans with fixed orientation such that +Y axis and +Z axis faces top and front/forward direction respectively and Z-axis reading is processed.

Accelerometer raw readings are saved in the .txt format so that it can be processed or cross-verified on the computer later, acting as a data logger.

The raw data is filtered by a low-pass(LP) filter.

LP filtered data passes via moving average filtered data x(default 5) times with a window size of x(default 5) to obtain a smooth graph.

Then the data is normalized.

Points of local maxima are recorded, and then points with a value less than a threshold value is discarded for each leg data set, thus giving us gait cycle points.

Average Gait Cycle(AGC) is calculated for each leg separately, that is the gait cycle template which has the closest resemblance to other gait cycles using dynamic time wrapping(dtw).

Average Gait Cycle Template calculated in previous step for each leg is compared using dynamic time wrapping(DTW), the value obtained is reported.