Survey on data fusion approaches for fall-detection

Elderly falls are a growing phenomenon observed within the world. According to World Health Organization (WHO), it is the second leading cause of unintentional or accidental deaths among the elderly. Thus, the need for research regarding the development of fall detection systems is imperative. Researchers have utilized various approaches to develop fall detection systems, significant number of which have employed Machine Leaning (ML) algorithms for fall detection. In this study, we evaluated the efficiency of six ML algorithms on a public fall detection dataset. A robust deep neural network for fall detection (FD-DNN) is identified to be the current state-of-the-art, it detects falls by using a self-built sensor that consumes low power. By evaluating the efficiency of six machine learning algorithms on a publicly available joint fall detection dataset, the accuracy of the fall detection was increased from 99.17% to 99.88% by using the K-nearest Neighbor indicating that common machine learning algorithms can achieve identical or higher accuracy rendering the complex and expensive deep neural network-based fall detection systems inefficient.KeywordsFall detectionElderly fall detectionMachine learning algorithmsWearable fall detection systemSisFall datasetMobiFall datasetK-nearest neighborsKNNSupport vector machinesSVMDecision treeRandom forestLogistic regressionNaïve Bayes

Falls are the main cause of injury in the elderly people. Therefore, fall detection system in the elderly is an essential technology for aging society. This project was designed to develop the fall detection system based on wearable sensors and Internet of Things for the real-time fall detection, data storage, calculation, and notification. The raw data, kinematics signal from the daily living of elder, including three-dimensional accelerometer and gyroscope are provided by GY-85, which have been sent bidots cloud via NodeMCU-ESP32. Further, the NodeMCU-ESP32 avail of the raw data to specify the fall detection algorithm. Moreover, the NodeMCU-ESP32 is designed to achieve the seamless communication between the fall detection device and system to acquire and process the data anywhere anytime as long as there is wifi connection. Consequently, the Ubidots cloud is designed to perform the calculation based on support vector machine (SVM) theory. It also sends the fall detection result via LINE notification (chat application) to caregivers. The tests of the proposed system are performed based on activities of daily living and the directions that fall can occur. The results of the testing show that the proposed system can provide high accuracy and reliability in fall detection.

Older adults face increased health risks, especially accidental injuries such as falls, which are one of the top ten causes of death in this age group. To address these changes, recent innovations have focused on developing advanced monitoring technologies to detect and prevent accidents in real time. Among these, fall detection systems have emerged as a critical area of research. This study aimed to evaluate the ability of millimeter-wave (mmWave) sensors to accurately detect multiple falls in the presence of large obstacles in a large, real-world indoor space. The mmWave sensors employed the Doppler effect to capture a human body’s point cloud and track its center point to estimate body position and identify fall events. A 12 $$\times$$ 12 meter indoor test area was established for the trials. The mmWave system’s accuracy was validated with video ground truth. Multiple sensors and azimuth tests were conducted to optimize radar configurations. 10 participants performed multiple human fall detection trials under 10 different scenarios. We have successfully validated the mmWave system with the video ground truth. In fall detection testing, the mmWave system achieved an overall accuracy of 97.9% across 10 multi-person scenarios. The results show that the system’s fall detection false negative rate increases with the number of subjects. This study validated the performance of a mmWave system for fall detection in a large indoor environment, demonstrating a high accuracy of 97.9% in a multi-person scenario. However, performance varied with crowd density, showing a correlation between increased false negative rates and the number of subjects due to occlusion. This study supports that mmWave technology offers good capabilities for fall accident monitoring in large indoor spaces with both privacy protection and convenience. IRB Registry: Institutional Review Board of the Chang Gung Medical Foundation, Approval No.: 202500191B0, Registration date: 3 March 2025.

Currently, there are several fall detection systems based on video analysis. However, these systems have not yet reached the desired level of appropriateness and robustness. To reduce the risk of falling in insecure environments, a new method is developed in this paper to detect and predict human fall detection. We adopt, in this approach, a Block Matching motion estimation algorithm based on acceleration and changes of the human body silhouette area, which are obtained from a single surveillance camera. It presents an algorithm to accelerate the fall detection system on based on a local adjustment of the velocity field.

Falling is among the most damaging event elderly people may experience. With the ever-growing aging population, there is an urgent need for the development of fall detection systems. Thanks to the rapid development of sensor networks and the Internet of Things (IoT), human-computer interaction using sensor fusion has been regarded as an effective method to address the problem of fall detection. In this paper, we provide a literature survey of work conducted on elderly fall detection using sensor networks and IoT. Although there are various existing studies which focus on the fall detection with individual sensors, such as wearable ones and depth cameras, the performance of these systems are still not satisfying as they suffer mostly from high false alarms. Literature shows that fusing the signals of different sensors could result in higher accuracy and lower false alarms, while improving the robustness of such systems. We approach this survey from different perspectives, including data collection, data transmission, sensor fusion, data analysis, security, and privacy. We also review the benchmark data sets available that have been used to quantify the performance of the proposed methods. The survey is meant to provide researchers in the field of elderly fall detection using sensor networks with a summary of progress achieved up to date and to identify areas where further effort would be beneficial.

Accidental falls are a major source of loss of autonomy, deaths, and injuries among the elderly. Accidental falls also have a remarkable impact on the costs of national health systems. Thus, extensive research and development of fall detection and rescue systems are a necessity. Technologies related to fall detection should be reliable and effective to ensure a proper response. This article provides a comprehensive review on state-of-the-art fall detection technologies considering the most powerful deep learning methodologies. We reviewed the most recent and effective deep learning methods for fall detection and categorized them into three categories: Convolutional Neural Network (CNN) based systems, Long Short-Term Memory (LSTM) based systems, and Auto-encoder based systems. Among the reviewed systems, three dimensional (3D) CNN, CNN with 10-fold cross-validation, LSTM with CNN based systems performed the best in terms of accuracy, sensitivity, specificity, etc. The reviewed systems were compared based on their working principles, used deep learning methods, used datasets, performance metrics, etc. This review is aimed at presenting a summary and comparison of existing state-of-the-art deep learning based fall detection systems to facilitate future development in this field.

The machine learning revolution is redesigning modern healthcare, and with the growth of the elderly population, fall detection has become an important research topic in healthcare. This paper surveys advances in machine learning-based fall detection technologies and reviews sensor-based, image processing-based, and wearable sensor-based fall detection systems and applications. In addition, this paper proposes a self-learning posture analysis and eye status-based fall detection system to solve the issue of mis-detections in fall detection systems, which have not been addressed in past works. Furthermore, this work proposes an image-feature-separation system that can use image processing with a low risk of privacy disclosure. Moreover, this work establishes a dataset, which includes 36 non-fall/fall cases comprising 25,200 images that can be used not only for this research but also in related studies. Experimental results show that this system can detect a fall with high accuracy and solve mis-detections in machine learning-based fall detection systems.

In the context of the ageing global population, researchers and scientists have tried to find solutions to many challenges faced by older people. Falls, the leading cause of injury among elderly, are usually severe enough to require immediate medical attention; thus, their detection is of primary importance. To this effect, many fall detection systems that utilize wearable and ambient sensors have been proposed. In this study, we compare three newly proposed data fusion schemes that have been applied in human activity recognition and fall detection. Furthermore, these algorithms are compared to our recent work regarding fall detection in which only one type of sensor is used. The results show that fusion algorithms differ in their performance, whereas a machine learning strategy should be preferred. In conclusion, the methods presented and the comparison of their performance provide useful insights into the problem of fall detection.

Falling can have fatal consequences for elderly people especially if the fallen person is unable to call for help due to loss of consciousness or any injury. Automatic fall detection systems can assist through prompt fall alarms and by minimizing the fear of falling when living independently at home. Existing vision-based fall detection systems lack generalization to unseen environments due to challenges such as variations in physical appearances, different camera viewpoints, occlusions, and background clutter. In this paper, we explore ways to overcome the above challenges and present Single Shot Human Fall Detector (SSHFD), a deep learning based framework for automatic fall detection from a single image. This is achieved through two key innovations. First, we present a human pose based fall representation which is invariant to appearance characteristics. Second, we present neural network models for 3d pose estimation and fall recognition which are resilient to missing joints due to occluded body parts. Experiments on public fall datasets show that our framework successfully transfers knowledge of 3d pose estimation and fall recognition learnt purely from synthetic data to unseen real-world data, showcasing its generalization capability for accurate fall detection in real-world scenarios.

Elderly are the world’s largest growing population, categorized over the age of 60 to 65 years. They are the ones who prone to fall due to their old age and low self-efficacy, thus making them vulnerable to different accidents. Even doing daily activities can also expose the elderly to a fall incident. As a result, it has gained the attention of many researchers in conducting studies related to the elderly daily health care, especially in relation to the fall detection system. This paper aims to provide a systematic review on the classification of fall detection systems for the elderly. This systematic review is designed based on the existing and extensive literature review on fall detection systems guided by the prisma statement (preferred reporting items for systematic reviews and meta-analyses) review method. Based on this systematic review, four overarching themes that provide in-depth information on fall detection to detect fall events have been identified; classification of fall detection, basis development, type of sensor and detection technique. In a nutshell, the fall detection approach has successfully provided an alternative health care services for elderly who choose to live independently. Therefore, it is important to continue to develop a fall detection system that integrates with technology in order to provide a safe living environment for elderly, and for children, it can offer as an alternative for monitoring systems.

Falls pose a significant risk for the elderly population, often resulting in serious injuries and decreased quality of life. Accurate and effective fall detection systems can play an important role in reducing these risks. This study conducts a comparative analysis of the performance of PySpark and Scikit-Learn libraries in the development of fall detection models. PySpark offers a distributed computing environment specifically designed for big data processing, while Scikit-Learn is a Python-based library widely used for data analysis and machine learning. Five popular machine learning algorithms, including logistic regression, gradient boosting classifier, random forest, support vector machine, and Decision Tree, were used to build fall detection models using both PySpark and Scikit-Learn. Models were evaluated using a comprehensive set of measurements including accuracy, sensitivity, specificity, and confusion matrix. In addition to the five popular machine learning algorithms included in the study, 26 different features were extracted from the Sisfall dataset, one of the most comprehensive fall and daily life activity datasets, in five main categories: basic statistical features, frequency domain features, time series features, motion features, and relational features. These features were incorporated into fall detection models to improve their ability to accurately identify falls. The findings reveal that both PySpark and Scikit-Learn provide strong and effective results in fall detection. Logistic regression emerged as the best performing algorithm in both libraries, achieving the highest accuracy rates. However, PySpark exhibited slightly longer training times compared to Scikit-Learn, which performed better in testing. Ultimately, this study will benefit researchers and practitioners involved in the development of robust and effective fall detection systems, thereby enabling improved safety and well-being for the elderly population. Additionally, it contributes to the literature by presenting a comprehensive feature extraction methodology for fall detection, which has not been explored extensively in previous research works.

Advances in modern medicine has increased humans’ life span. Orderly adults face mobility problems while aging. They feel less fit to continue any activity for short intervals too. This is due to declining fitness levels or muscle strength, diminished dexterity, and loss of balance. These symptoms lead to the fall of the individual and sometimes fatal too, if immediately not attended to. It’s an alarming issue for people staying alone. They may pose significant health risks and need immediate assistance. Fall detection technologies are majorly categorised as wearable sensors and ambient sensors. Fall detection wearable devices like pendant necklaces, watches and wristband devices, and clip-on medical alerts use accelerometers to detect rapid downward movements that can indicate a fall. They often also include manual alert buttons, for an increased accuracy. This requires technology comfort and awareness for usage. Ambient home sensors use video cameras to monitor the user’s movement and detect falls. When the fall is transmitted to a monitoring center, a representative typically will call the user to check on them before notifying contacts or calling for emergency services, but this can depend on the user’s preferences and risk factors. In this paper we propose a technology, using security cameras to record videos and create a video-based fall detection system. The system uses computer vision and deep learning algorithms to accurately recognize fall-related movements and distinguish them from regular activities. This system can be integrated to prompt alerts to emergency contacts, thus assisting in providing immediate aid to individuals who have experienced a fall. For higher accuracy, multiple-angle videos and multi-person tracking is integrated in this system to estimate the intensity of the fall for immediate attention. Thus, this fall detection system can contribute to the safety, well-being and independence of individuals at risk of falling.

Falls among older people remain a very important public healthcare issue. Every year over 11 million falls are registered in the U.S. alone. This paper presents a practical real time fall detection system running on a smartwatch (F2D). A decision module takes into account the rebound after the fall and the residual movement of the user, matching a detected fall pattern to an actual fall. The final decision of a fall event is taken based on the location of the user. To the best of our knowledge, this is the first fall detection system which works on an independent smartwatch, being less stigmatizing for the end user. The fall detection algorithm has been tested by Fondation Suisse pour les Téléthèses (FST), the project partner who is responsible for the commercialization of our system. By analyzing real data of activities of daily life of elderly people, we are confident that F2D meets the demands of a reliable and easily extensible system. This paper highlights the innovative algorithm which takes into account the residual movement and the location of the user to increase the fall detection accuracy. By testing with real data we have a fall detection system ready to be deployed on the market.

Falls among older adults pose significant health risks, making their prevention and detection critical areas of research. This review examines fall detection and prevention systems, categorizing them based on sensor types and utilization methods: wearable sensors, environmental sensors, radio-frequency-based sensors, and hybrid systems. Additionally, it explores the methods employed within these systems. Given the limitations of traditional linear approaches in accurately detecting falls, recent research emphasizes artificial intelligence (AI) techniques, particularly machine learning (ML) and deep learning (DL), to enhance detection accuracy and system functionality. The review provides an overview of the sensors and algorithms used in fall detection and prevention systems, alongside their outcomes. Key findings and challenges related to specific sensors and systems are discussed in detail. This analysis offers researchers a comprehensive understanding of current technologies, highlights the contributions of AI methods, and outlines potential future directions in the field. By evaluating sensors, methodologies, and system sensitivities, the aim is to contribute to the development of effective solutions tailored to specific sensitivities.

Ambient assisted living is a concept which uses information and communication technology to assist the daily living of people. Human fall detection is an important sub-area of ambient assisted living. Human fall has been seen as a critical problem for elderly people. Fall detection is an approach which analyzes sensor data (wearable sensors/ambient sensors or vision-based sensor) to detect human fall using various learning algorithms. This paper presents a fall detection method that detects and notifies fall activity in real-time using fog computing. Support Vector Machine based one class classification is used here to build fall detection model. Five features have been calculated from Smartphone accelerometer data to build fall detection model. To implement one class classification, a new method for kernel matrix calculation is proposed here. This fall detection model exploited the concept of fog computing to send real-time notification to the caregiver and it is also able to notify caregiver in absence of fog node to cloud connection. In the proposed method we have got 100% sensitivity and 98.77% specificity for human fall detection. This fall detection method is also tested on real fall data and it is found that this method is able to detect 100% fall activities. Use of fog computing concept drastically reduces amount of data transferred to the cloud from 900 values (10,799 bytes) to 5 values (59 bytes) per 6 s.