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

Smartphone-Based Analysis for Early Detection of Aging Impact on Gait and Stair Negotiation: A Cross-Sectional Study

Aging is associated with gradual mobility decline, often undetected until it affects daily life. This study investigates the potential of smartphone-based accelerometry to detect early age-related changes in gait and stair performance in middle-aged adults. Eighty-eight healthy participants were divided into four age groups: young (20-35 years), early middle-aged (45-54 years), late middle-aged (55-65 years), and older adults (65-80 years). They completed single-task, cognitive, and physical dual-task gait assessments and stair negotiation tests. While single-task walking did not reveal early changes, cognitive dual-task cost (DTC) of stride time variability deteriorated in late middle age. A strong indicator of early mobility changes was movement similarity, measured using dynamic time warping (DTW), which declined from early middle age for both cognitive DTC and stair negotiation. These findings highlight the potential of smartphone-based assessments, particularly movement similarity, to detect subtle mobility changes in midlife, allowing for targeted interventions to promote healthy aging.

High-Knee-Flexion Posture Recognition Using Multi-Dimensional Dynamic Time Warping on Inertial Sensor Data

Relating continuously collected inertial data to the activities or postures performed by the sensor wearer requires pattern recognition or machine-learning-based algorithms, accounting for the temporal and scale variability present in human movements. The objective of this study was to develop a sensor-based framework for the detection and measurement of high-flexion postures frequently adopted in occupational settings. IMU-based joint angle estimates for the ankle, knee, and hip were time and scale normalized prior to being input to a multi-dimensional Dynamic Time Warping (mDTW) distance-based Nearest Neighbour algorithm for the identification of twelve postures. Data from 50 participants were divided to develop and evaluate the mDTW model. Overall accuracies of 82.3% and 55.6% were reached when classifying movements from the testing and validation datasets, respectively, which increased to 86% and 74.6% when adjusting for imbalances between classification groups. The highest misclassification rates occurred between flatfoot squatting, heels-up squatting, and stooping, while the model was incapable of identifying sequences of walking based on a single stride template. The developed mDTW model proved robust in identifying high-flexion postures performed by participants both included and precluded from algorithm development, indicating its strong potential for the quantitative measurement of postural adoption in real-world settings.

Examining sensor-based physical activity recognition and monitoring for healthcare using Internet of Things: A systematic review

Due to importantly beneficial effects on physical and mental health and strong association with many rehabilitation programs, Physical Activity Recognition and Monitoring (PARM) have been considered as a key paradigm for smart healthcare. Traditional methods for PARM focus on controlled environments with the aim of increasing the types of identifiable activity subjects complete and improving recognition accuracy and system robustness by means of novel body-worn sensors or advanced learning algorithms. The emergence of the Internet of Things (IoT) enabling technology is transferring PARM studies to open and connected uncontrolled environments by connecting heterogeneous cost-effective wearable devices and mobile apps. Little is currently known about whether traditional PARM technologies can tackle the new challenges of IoT environments and how to effectively harness and improve these technologies. In an effort to understand the use of IoT technologies in PARM studies, this paper will give a systematic review, critically examining PARM studies from a typical IoT layer-based perspective. It will firstly summarize the state-of-the-art in traditional PARM methodologies as used in the healthcare domain, including sensory, feature extraction and recognition techniques. The paper goes on to identify some new research trends and challenges of PARM studies in the IoT environments, and discusses some key enabling techniques for tackling them. Finally, this paper consider some of the successful case studies in the area and look at the possible future industrial applications of PARM in smart healthcare.

Automated classification of pathological gait after stroke using ubiquitous sensing technology

This study uses machine learning methods to distinguish between healthy and pathological gait. Examples of multi-dimensional pathological and normal gait sequences were collected from post-stroke and healthy individuals in a real clinical setting and with two Kinect sensors. The trajectories of rotational angle and global velocity of selected body joints (hips, spine, shoulders, neck, knees and ankles) over time formed the gait sequences. The combination of k nearest neighbor (kNN) and dynamic time warping (DTW) was used for classification. Leave one subject out cross validation was implemented to evaluate the performance of the binary classifier in terms of F1-score in the original feature space, and also in a reduced dimensional feature space using PCA. The pair of k = 1 in kNN and the warping window size 25% of gait sequences in DTW achieved maximum F1-score. Using PCA, pathological gait sequences were discriminated from healthy sequences with the F1-score = 96%.

Humanoid assessing rehabilitative exercises

INTRODUCTION

This article is part of the Focus Theme of Methods of Information in Medicine on "New Methodologies for Patients Rehabilitation".

BACKGROUND

The article presents the approach in which the rehabilitative exercise prepared by healthcare professional is encoded as formal knowledge and used by humanoid robot to assist patients without involving other care actors.

OBJECTIVES

The main objective is the use of humanoids in rehabilitative care. An example is pulmonary rehabilitation in COPD patients. Another goal is the automated judgment functionality to determine how the rehabilitation exercise matches the pre-programmed correct sequence.

METHODS

We use the Aldebaran Robotics' NAO humanoid to set up artificial cognitive application. Pre-programmed NAO induces elderly patient to undertake humanoid-driven rehabilitation exercise, but needs to evaluate the human actions against the correct template. Patient is observed using NAO's eyes. We use the Microsoft Kinect SDK to extract motion path from the humanoid's recorded video. We compare human- and humanoid-operated process sequences by using the Dynamic Time Warping (DTW) and test the prototype.

RESULTS

This artificial cognitive software showcases the use of DTW algorithm to enable humanoids to judge in near real-time about the correctness of rehabilitative exercises performed by patients following the robot's indications.

CONCLUSION

One could enable better sustainable rehabilitative care services in remote residential settings by combining intelligent applications piloting humanoids with the DTW pattern matching algorithm applied at run time to compare humanoid- and human-operated process sequences. In turn, it will lower the need of human care.

Comparative evaluation of features and techniques for identifying activity type and estimating energy cost from accelerometer data

Wearable accelerometers can be used to objectively assess physical activity. However, the accuracy of this assessment depends on the underlying method used to process the time series data obtained from accelerometers. Several methods have been proposed that use this data to identify the type of physical activity and estimate its energy cost. Most of the newer methods employ some machine learning technique along with suitable features to represent the time series data. This paper experimentally compares several of these techniques and features on a large dataset of 146 subjects doing eight different physical activities wearing an accelerometer on the hip. Besides features based on statistics, distance based features and simple discrete features straight from the time series were also evaluated. On the physical activity type identification task, the results show that using more features significantly improve results. Choice of machine learning technique was also found to be important. However, on the energy cost estimation task, choice of features and machine learning technique were found to be less influential. On that task, separate energy cost estimation models trained specifically for each type of physical activity were found to be more accurate than a single model trained for all types of physical activities.

User-Independent Recognition of Sports Activities From a Single Wrist-Worn Accelerometer: A Template-Matching-Based Approach

GOAL

To investigate the accuracy of template matching for classifying sports activities using the acceleration signal recorded with a wearable sensor.

METHODS

A population of 29 normal weight and 19 overweight subjects was recruited to perform eight common sports activities, while body movement was measured using a triaxial accelerometer placed at the wrist. User- and axis-independent acceleration signal templates were automatically extracted to represent each activity category and recognize activity types. Five different similarity measures between example signals and templates were compared: Euclidean distance, dynamic time warping (DTW), derivative DTW, correlation and an innovative index, and combining distance and correlation metrics ( Rce). Template-based activity recognition was compared to statistical-learning classifiers, such as Naïve Bayes, decision tree, logistic regression (LR), and artificial neural network (ANN) trained using time- and frequency-domain signal features. Each algorithm was tested on data from a holdout group of 15 normal weight and 19 overweight subjects.

RESULTS

The Rce index outperformed other template-matching metrics by achieving recognition rate above 80% for the majority of the activities. Template matching showed robust classification accuracy when tested on unseen data and in case of limited training examples. LR and ANN achieved the highest overall recognition accuracy  ∼  85% but showed to be more vulnerable to misclassification error than template matching on overweight subjects' data.

CONCLUSION

Template matching can be used to classify sports activities using the wrist acceleration signal.

SIGNIFICANCE

Automatically extracted template prototypes from the acceleration signal may be used to enhance accuracy and generalization properties of statistical-learning classifiers.

SVM versus MAP on accelerometer data to distinguish among locomotor activities executed at different speeds

Two approaches to the classification of different locomotor activities performed at various speeds are here presented and evaluated: a maximum a posteriori (MAP) Bayes' classification scheme and a Support Vector Machine (SVM) are applied on a 2D projection of 16 features extracted from accelerometer data. The locomotor activities (level walking, stair climbing, and stair descending) were recorded by an inertial sensor placed on the shank (preferred leg), performed in a natural indoor-outdoor scenario by 10 healthy young adults (age 25-35 yrs.). From each segmented activity epoch, sixteen features were chosen in the frequency and time domain. Dimension reduction was then performed through 2D Sammon's mapping. An Artificial Neural Network (ANN) was trained to mimic Sammon's mapping on the whole dataset. In the Bayes' approach, the two features were then fed to a Bayes' classifier that incorporates an update rule, while, in the SVM scheme, the ANN was considered as the kernel function of the classifier. Bayes' approach performed slightly better than SVM on both the training set (91.4% versus 90.7%) and the testing set (84.2% versus 76.0%), favoring the proposed Bayes' scheme as more suitable than the proposed SVM in distinguishing among the different monitored activities.

LocateMe

Fine-grained localization is extremely important to accurately locate a user indoors. Although innovative solutions have already been proposed, there is no solution that is universally accepted, easily implemented, user centric, and, most importantly, works in the absence of GSM coverage or WiFi availability. The advent of sensor rich smartphones has paved a way to develop a solution that can cater to these requirements.

By employing a smartphone's built-in magnetic field sensor, magnetic signatures were collected inside buildings. These signatures displayed a uniqueness in their patterns due to the presence of different kinds of pillars, doors, elevators, etc., that consist of ferromagnetic materials like steel or iron. We theoretically analyze the cause of this uniqueness and then present an indoor localization solution by classifying signatures based on their patterns. However, to account for user walking speed variations so as to provide an application usable to a variety of users, we follow a dynamic time-warping-based approach that is known to work on similar signals irrespective of their variations in the time axis.

Our approach resulted in localization distances of approximately 2m--6m with accuracies between 80--100% implying that it is sufficient to walk short distances across hallways to be located by the smartphone. The implementation of the application on different smartphones yielded response times of less than five secs, thereby validating the feasibility of our approach and making it a viable solution.

Impact of Sensor Misplacement on Dynamic Time Warping Based Human Activity Recognition using Wearable Computers

Daily living activity monitoring is important for early detection of the onset of many diseases and for improving quality of life especially in elderly. A wireless wearable network of inertial sensor nodes can be used to observe daily motions. Continuous stream of data generated by these sensor networks can be used to recognize the movements of interest. Dynamic Time Warping (DTW) is a widely used signal processing method for time-series pattern matching because of its robustness to variations in time and speed as opposed to other template matching methods. Despite this flexibility, for the application of activity recognition, DTW can only find the similarity between the template of a movement and the incoming samples, when the location and orientation of the sensor remains unchanged. Due to this restriction, small sensor misplacements can lead to a decrease in the classification accuracy. In this work, we adopt DTW distance as a feature for real-time detection of human daily activities like sit to stand in the presence of sensor misplacement. To measure this performance of DTW, we need to create a large number of sensor configurations while the sensors are rotated or misplaced. Creating a large number of closely spaced sensors is impractical. To address this problem, we use the marker based optical motion capture system and generate simulated inertial sensor data for different locations and orientations on the body. We study the performance of the DTW under these conditions to determine the worst-case sensor location variations that the algorithm can accommodate.