Rossana Muscillo

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

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).

Early recognition of upper limb motor tasks through accelerometers: real-time implementation of a DTW-based algorithm

A new real-time implementation of a Dynamic Time Warping (DTW)-based classification scheme is presented here, and its performance evaluated on experimental data. Nine young adults were requested to perform instances of eight different purposeful movements described in the Wolf Motor Function Test, while wearing a three-axis accelerometer sensor placed on the inner forearm. Results include the correct recognition percentage, as compared to a classification scheme based on the traditional DTW measure, and the recognition percentage as a function of the time elapsed from the beginning of the performed movements. The Real-Time DTW basically performs with the same accuracy of the traditional DTW-based classification scheme (91.5% of correct recognition percentage), a figure that increases to 96.5% if the multidimensional scheme is adopted. Moreover, more than 60% of movements are correctly recognized before their end, thus setting the way for applications in rehabilitation and assistive technologies, where a real-time control scheme is able to interact with the user while the movement is being performed.

An adaptive Kalman-based Bayes estimation technique to classify locomotor activities in young and elderly adults through accelerometers

An accelerometer-based system able to classify among different locomotor activities during real life conditions is here presented, and its performance evaluated. Epochs of walking at different speeds, and with different slopes, and stair descending and ascending, are detected, segmented, and classified by using an adaptation of a naïve 2D-Bayes classifier, which is updated on-line through the history of the estimated activities, in a Kalman-based scheme. The feature pair used for classification is mapped from an ensemble of 16 features extracted from the accelerometer data for each activity epoch. Two different versions of the classifier are presented to combine the multi-dimensional nature of the accelerometer data, and their results are compared in terms of correct recognition rate of the segmented activities, on two population samples of different age. The classification algorithm achieves correct classification rates higher than 90% and higher than 92%, for young and elderly adults, respectively.

The Median Point DTW Template to Classify Upper Limb Gestures at Different Speeds

Detecting activities of daily living (ADL) and classifying the gesture typology are important tasks for rehabilitation and for applications in robotics. The use of wearable sensors, such as accelerometers, could facilitate the previous tasks since it would open the possibility of monitoring patients in real-life conditions.

Estimation of Physical Performance for Elderly People via Markerless Sit to Stand Analysis

We propose a novel system for automatically assessing physical performance in elderly people. We concentrated on the kinematic analysis of the Sit-to-Stand (STS) task since it is widely recognised that the inability to rise from a seated position is associated with physical impairment and risk of falling. The approach implements a markerless kinematic analysis by applying a method working on the Gauss-Laguerre transformed domain. Limbs’ pose is analysed and compared with a normality trend, obtained from young subjects. A parameter representing the strategy used for raising from the chair, and correlated to the physical assessment, is extracted. Experimental tests on a population of elderly people are conducted and results are compared with scores from the Short Physical Performance Battery and the Tinetti’s Test.

Real-time adaptive neural predictors for upper limb gestures blind recognition

A real-time system based on artificial neural networks (ANNs) for time series prediction is proposed. For each movement the prediction errors are used both to train the ANNs and to estimate a measure of the unlikelihood of the specific gesture occurrence. The first repetition of each ges- ture trains the related ANNs bank and the current motion is recognized after a few successive repetitions. Neither a priori assumptions nor signal pre-processing is performed (blind- ness). The training is performed at the beginning and can be repeated during the running (adaptability). Each procedure is strictly physically realizable (real-time feasibility). Four gestures performed by three healthy volunteers are se- lected by a set of upper limb motor tasks contained in the rehabilitation scale known as Wolf motor function test (WMFT). Two accelerometers placed on the upper arm and on the forearm respectively, constitute the sensors set. Even in very challenging test settings, the proposed method shows a correct recognition rate higher than 83%.