Search

In this paper, a new method is proposed for people tracking using the human skeleton provided by the Kinect sensor, Our method is based on skeleton data, which includes the coordinate value of each joint in the human body. For data classification, the Support Vector Machine (SVM) and Random Forest techniques are used. To achieve this goal, 14 classes of movements are defined, using the Kinect Sensor to extract data containing 46 features and then using them to train the classification models. The system was tested on 12 subjects, each of whom performed 14 movements in each experiment. Experiment results show that the best average accuracy is 90.2 % for the SVM model and 99 % for the Random forest model. From the experiments, we concluded that the best distance between the Kinect sensor and the human body is one meter.

In-bed motion detection and classification are important techniques that can enable an array of applications, among which are sleep monitoring and abnormal movement detection. In this article, we present a low-cost, low-overhead, and highly robust system for in-bed movement detection and classification that uses low-end load cells. To detect movements, we have designed a feature that we refer to as Log-Peak, which can be extracted from load cell data that is collected through wireless links in an energy-efficient manner. After detection, we set out to achieve a precise body motion classification. Toward this goal, we define nine classes of movements, and design a machine learning algorithm using Support Vector Machine, Random Forest, and XGBoost techniques to classify a movement into one of nine classes. For every movement, we have extracted 24 features and used them in our model. This movement detection/classification system was evaluated on data collected from 40 subjects who performed 35 predefined movements in each experiment. We have applied multiple tree topologies for each technique to reach their best results. After examining various combinations, we have achieved a final classification accuracy of 91.5%. This system can be used conveniently for long-term home monitoring.

Medication compliance is challenging to monitor and identify because of their complex nature that is significantly influenced by the patient's behavior. Medication compliance monitoring systems are fundamental contributions to the connected health paradigm. In this paper, we propose a compliance sensing framework that does not require on-body sensing modalities or costly installation and maintenance. It uses a smart pill-bottle that utilizes Inertial Measurement Units (IMUs) and machine learning approaches for activity classification. The IMUs collect data about user behavior while taking medication. The collected signals are subject to pre-processing, from which features can be extracted for activity detection algorithms. Our initial results show promise in correctly discriminating medication-taking 90 % of the time compare to only moving the bottle.

Measuring in-bed mobility is a very significant consideration when it comes to tracking patients or individuals during sleep. A variety of applications, such as sleep monitoring and unusual movements during sleep, can be enabled by observing a persons body movements throughout sleep. In-bed movement can be a sign of sleep disruption as it is associated with wakefulness that affects the quality of sleep and can be a sign of many illnesses. We introduce, in this study, an unobtrusive scheme for in-bed motion detection and classification using geophones sensor. Geophone can sense the vibration that caused by every in-bed movement. We have extracted two features from the sensed signal, which we named as Energy-Peak and Log-Peak. We, at that point, utilized a straightforward threshold-based calculation to identify each conceivable movement. In addition to movements detection, we further classify them as a big or small motion. We have assessed this framework by doing 30 tests with 15 members over a two-month. There are 35 movements in each experiment. By using our two primary approaches, Energy-Peak and Log-Peak, our system can identify in-bed motions with a 2% as a low error rate. For classification phase, we have extracted 4 features from every detected movement and we have used Random Forest technique for classification decision. Our system can classify every movement as big or small with 1.5% as an error rate.

The basic necessity of sleep in our life is critically important to ensure our wellbeing. Sufficient sleep of good quality is highly desired in order to have enough energy to live. One of the main factors to measure sleep quality is the amount of body motion during sleep. In our earlier work, we have developed a load cell based system that can detect in-bed body movements and classify them into two broad classes: large or small movements. In this paper, we set out to achieve much finer body motion classification. Towards this goal, we define 9 classes of movements, and design a machine learning algorithm using Support Vector Machine (SVM) and Random Forest techniques to classify a movement into one of these 9 classes. In this way, we can find out which body parts are involved in every movement. For every movement, we have extracted 24 features and used them in our model. This movement classification system was evaluated on data collected from 40 subjects who performed 35 predefined movements in each experiment. The accuracy of our model is not exactly the same for all classes of movements. On average it classified correctly 90% of movements. This model can be used conveniently for long-term home monitoring.

In-bed motion detection is an important techniquethat can enable an array of applications, among which aresleep monitoring and abnormal movement detection. In thispaper, we present a low-cost, low-overhead, and highly robustsystem for in-bed movement detection and classification thatuses low-end load cells. By observing the forces sensed by theload cells, placed under each bed leg, we can detect manydifferent types of movements, and further classify them as bigor small depending on magnitude of the force changes on theload cells. We have designed three different features, whichwe refer to as Log-Peak, Energy-Peak, ZeroX-Valley, that caneffectively extract body movement signals from load cell datathat are collected through wireless links in an energy-efficientmanner. After establishing the feature values, we employ a simplethreshold-based algorithm to detect and classify movements. Wehave conducted thorough evaluation, that involves collecting datafrom 30 subjects who perform 27 pre-defined movements in anexperiment. By comparing our detection and classification resultsagainst the ground truth captured by a video camera, we showthe Log-Peak strategy can detect these 27 types of movementsat an error rate of 6.3% while classifying them to big or smallmovements at an error rate of 4.2%.