Fall Detection Analysis Using a Real Fall Dataset

2014

Falls and fall-related injuries are major incidents, especially for elderly people, which often mark the onset of major deterioration of health. More than one-third of home-dwelling people aged 65 or above and two-thirds of those in residential care fall once or more each year. Reliable fall detection, as well as prevention, is an important research topic for monitoring elderly living alone in residential or hospital units. The aim of this study is to review the existing fall detection systems and some of the key research challenges faced by the research community in this field. We categorize the existing platforms into two groups: wearable and ambient devices; the classification methods are divided into rule-based and machine learning techniques. The relative merit and potential drawbacks are discussed, and we also outline some of the outstanding research challenges that emerging new platforms need to address.

Proceedings of the International Conference on Biomedical Electronics and Devices, 2012

As in previous editions of the Conference, based on the reviewers' evaluations and on the quality of presentations, a short Iist of authors will be selected and invited to submit extended revised versions of their papers for a book that will be published by Springer with the best papers of BIOSTEC 2012. We would like to express our thanks to ali participants. First of ali, to the authors, whose quality work is the essence of this joint conference. Next, we thank ali the members of the program committee and the auxiliary reviewers for their diligence and expert reviewing. We would also like to deeply thank the invited speakers for their excellent contribution in sharing their knowledge and vision. Fourthly, we thank the program chairs, Joaquim Gabriel, Jan Schier, Sabine Van Huffel and Emmanuel Conchon whose collaboration was much appreciated. Finally, special thanks to ali the members of the INSTICC team whose collaboration was fundamental for lhe success of Ihis conference. We wish you ali an inspiring conference and an unforgettable stay in Vilamoura, Algarve and we hope to meet you again nexl year for the 6th BIOSTEC, details of which will soon be available at http://www.biostec.org.

Applied Soft Computing, 2014

This paper presents the development of a smart fall detector to minimise accidental falls which occur among elderly people, especially for indoor situations. A video-based detection system was utilised, as this can preserve privacy and monitor the physical activities of elderly people. In order to identify the correct situation among a set of predetermined situations, which consisted of praying, sitting, standing, bending, kneeling and lying down, a neural network system was incorporated in the fall detection computation algorithm. The neural network analysed the binary map image of the person and then identified which plausible situation the person was in at any particular instant in time. The fall detector's performance in successfully detecting falls was then evaluated using two statistical metrics: specificity and sensitivity. The performance of this fall detection system in identifying falls was also evaluated during two nonnormal gait movements, stumbling and limping, so as to mimic the motions of a good proportion of the elderly people having these types of walking gait movements. It was shown that the implemented video-based fall detection system could be a promising solution for detecting indoor falls among senior citizens.

International Journal of Integrated Engineering, 2020

Falling is a well-known as a threat among elderly people which may lead to injuries or even death. High recognition of developed fall detection model is very significance for the elderly to detect the falls. Related algorithm for the fall detection has been discussed in depth by researcher during the past decade. However, the improvement of model accuracy is still needed. This article presents results of modelling for fall detection system by using nonlinear autoregression neural network NARnet algorithm. The algorithm is trained by network training function; LM, SCG and RP by collocation with threshold-based setting value. Two participants involved in obtaining the acceleration and angular velocity. The training is divided into 4 different type of input source. The selection of the model was based on the comparison of optimization epochs, magnitude of validate error or mean square error (MSE), magnitude of correlation performance, the progress of convergence graph in term of MSE pe...

Falling is a major problem among the people globally, frequently due to some health problems including vision loss or balance disorder as a consequence of aging. As a result of the falling in elder people; injuries, complications, neurological problems and mortality are generally occurred. This situation also affects the patients' families psychologically. A large number of studies show that significant fractures, injuries and in some cases death are commonly encountered in elderly. Nonetheless, in the immediate intervention, the rate of damage is reduced and the life quality can be significantly restored. The purpose of this paper is to reduce fall-related problems by developing a new fall detector that we called Tesodev fall detector. In addition, a proper algorithm and a wearable electronic device attaching the sensor to patient's clothes are provided. The electronic device is an IOT device which can send and receive data wirelessly, which means that the device can inform the medical centres to improve medical attention time. Also, this device is modular and device allows keeping other medical data such as blood sugar, tension, heart rate, SPO 2. Sensitivity, error rate and classification accuracy in this study are 89.8%, 23.4%, 76%, respectively.

Sensors, 2014

Falls and fall-related injuries are major incidents, especially for elderly people, which often mark the onset of major deterioration of health. More than one-third of home-dwelling people aged 65 or above and two-thirds of those in residential care fall once or more each year. Reliable fall detection, as well as prevention, is an important research topic for monitoring elderly living alone in residential or hospital units. The aim of this study is to review the existing fall detection systems and some of the key research challenges faced by the research community in this field. We categorize the existing platforms into two groups: wearable and ambient devices; the classification methods are divided into rule-based and machine learning techniques. The relative merit and potential drawbacks are discussed, and we also outline some of the outstanding research challenges that emerging new platforms need to address.

2020

Falls inside of the home is a major concern facing the aging population. Monitoring the home environment to detect a fall can prevent profound consequences due to delayed emergency response. One option to monitor a home environment is to use a camera-based fall detection system. Conceptual designs vary from 3D positional monitoring (multicamera monitoring) to body position and limb speed classification. Research shows varying degree of success with such concepts when designed with multi-camera setup. However, camera-based systems are inherently intrusive and costly to implement. In this research, we use a sound-based system to detect fall events. Acoustic sensors are used to monitor various sound events and feed a trained machine learning model that makes predictions of a fall events. Audio samples from the sensors are converted to frequency domain images using Mel-Frequency Cepstral Coefficients method. These images are used by a trained convolution neural network to predict a fall. A publicly available dataset of household sounds is used to train the model. Varying the model's complexity, we found an optimal architecture that achieves high performance while being computationally less extensive compared to the other models with similar performance. We deployed this model in a NVIDIA Jetson Nano Developer Kit.

SENSORDEVICES 2018, the Ninth International Conference on Sensor Device Technologies and Applications, Venice, Italy, 16-20 September 2018, 2018

Falling is a commonly occurring mishap with elderly people, which may cause serious injuries. Thus, rapid fall detection is very important in order to mitigate the severe effects of fall among the elderly people. Many fall monitoring systems based on the accelerometer have been proposed for the fall detection. However, many of them mistakenly identify the daily life activities as fall or fall as daily life activity. To this aim, an efficient machine learning-based fall detection algorithm has been proposed in this paper. The proposed algorithm detects fall with efficient sensitivity, specificity, and accuracy as compared to the state-of-the-art techniques. A publicly available dataset with a very simple and computationally efficient set of features is used to accurately detect the fall incident. The proposed algorithm reports and accuracy of 99.98% with the Support Vector Machine (SVM) classifier.

Procedia Computer Science, 2018

The elderly population is growing every year in Brazil. Consequently, health risks in elderly is a concern for public health system. During the aging process, the mobility is affected, and falls are more frequent causing injuries and even death, whose causes can be prevented, with reduction of financial costs. Therefore, a low-cost inertial sensor-based system is a tool to fulfill the need for detecting falls in elderly. In this paper, we present our system as a proof of concept for the study of fall and we propose a low cost and more accessible system for fall detection using inertial sensors. The inertial sensor collects data, identifies and detect four different "fall states". The aim is to use this system in public health. In real-time, it will advise any person around the elder about the fall. Different machine learning classifiers are tested in the train dataset, and the best one was used for training the sensor data. Then, the model was compared with unknown sensor data (captured and from available datasets) to guess at which state the person is. We found out that there were only 15 wrong observations from all trials, thus, the system has potential to be used to detect falls.

Digital Signal Processing, 2021

Effective fall-detection and classification systems are vital in mitigating severe medical and economical consequences of falls to people in the fall risk groups. One class of such systems is based on wearable sensors. While there is a vast amount of academic work on this class of systems, not much effort has been devoted to the investigation of effective and robust algorithms and like-for-like comparison of state-ofthe-art algorithms using a sufficiently large dataset. In this article, fall-direction classification algorithms are presented and compared on an extensive dataset, comprising a total of 1600 fall trials. Eight machine learning classifiers are implemented for fall-direction classification into four basic directions (forward, backward, right, and left). These are, namely, Bayesian decision making (BDM), least squares method (LSM), k-nearest neighbor classifier (k-NN), artificial neural networks (ANNs), support vector machines (SVMs), decision-tree classifier (DTC), random forest (RF), and adaptive boosting or AdaBoost (AB). BDM achieves perfect classification, followed by k-NN, SVM, and RF. Data acquired from only a single motion sensor unit, worn at the waist of the subject, are processed for experimental verification. Four of the classifiers (BDM, LSM, k-NN, and ANN) are modified to handle the presence of data from an unknown class and evaluated on the same dataset. In this robustness analysis, ANN and k-NN yield accuracies above 96.2%. The results obtained in this study are promising in developing real-world fall-classification systems as they enable fast and reliable classification of fall directions.