Classification of Motor Activities through Derivative Dynamic Time Warping applied on Accelerometer Data

Abstract

In the context of tele-monitoring, great interest is presently devoted to physical activity, mainly of elderly or people with disabilities. In this context, many researchers studied the recognition of activities of daily living by using accelerometers. The present work proposes a novel algorithm for activity recognition that considers the variability in movement speed, by using dynamic programming. This objective is realized by means of a matching and recognition technique that determines the distance between the signal input and a set of previously defined templates. Two different approaches are here presented, one based on Dynamic Time Warping (DTW) and the other based on the Derivative Dynamic Time Warping (DDTW). The algorithm was applied to the recognition of gait, climbing and descending stairs, using a biaxial accelerometer placed on the shin. The results on DDTW, obtained by using only one sensor channel on the shin showed an average recognition score of 95%, higher than the values obtained with DTW (around 85%). Both DTW and DDTW consistently show higher classification rate than classical Linear Time Warping (LTW).

Driving style recognition using a smartphone as a sensor platform

“…In recent years, many papers have surfaced involving gesture recognition using accelerometers [20], [21]. Devices such as the WiiMote are able to provide raw sensor data to PCs wirelessly, making them advantageous to research.…”

DTWDIR: An Enhanced DTW Algorithm for Autistic Child Behaviour Monitoring

“…Most of previous researches [6] [7] [11] used decision tree algorithm (C4.5) for recognizing stereotypical motor movements to identify the autistic child but Muscillo et al [19] displayed that DTW is better than C4.5 when they compared between them. As mentioned before, stereotypical motor movements are not enough to identify the autistic child therefore they were not in consideration, however we compared hand flapping detection accuracy in both cases when DTWDir achieved highest, and lowest accuracy using three axes together, and Z-Axis respectively.…”

“…The data are collected from two subjects -female, male-and they found that the mean accuracy of DTW is 91% even though the dataset training size is small. Muscillo et al [19] compared between decision tree C4.5 and DTW classifiers for recognizing low-level food preparation activities. The data are collected from four sensors equipped kitchen utensils (knifes and spoon) about 6 hours.…”

Coarse In-Building Localization with Smartphones

“…Dynamic time warping (DTW) is an algorithm for measuring similarity between two signals which may vary in time or speed [12]. It has been used successfully to classify activities that have a cyclic pattern.…”