A Comprehensive Review on Machine Learning Models for Real Time Fall Prediction Using Wearable Sensor-Based Gait Analysis

Falls are a significant health concern for the elderly population, and fall-related injuries can lead to severe consequences such as disability, loss of independence, and even death. Therefore, early fall prediction and prevention are crucial to ensure the well-being of elderly individuals. In recent years, the use of wearable sensors for gait analysis and fall prediction has gained significant attention in the research community. The purpose of this document is to provide a detailed overview of the latest advancements in real-time fall prediction using wearable sensors and gait analysis. A comprehensive review has been done on various studies that have employed different machine learning algorithms and sensor modalities to predict falls in real-time. Additionally, it describes the challenges and limitations associated with wearable sensor-based fall prediction, and it identifies the possible future research directions in this field and provides a comprehensive narrative review of the recent research on fall risk assessment using wearable sensors. It deliberates the various approaches and methodologies used for fall risk assessment and present an overview of the datasets and machine learning techniques employed for fall risk prediction and also highlight the challenges and limitations of wearable sensors for fall risk assessment and provide recommendations for future research in this area.