TinyML-based Real-Time System for Detecting Falls in Elderly People

Falls are one of the leading causes of death in elderly people. This can even lead to serious injuries or, in some cases, be fatal. The main issue is not just the fall itself but the lack of timely help afterwards, which causes more health problems. That is why having a system that can detect when a fall is crucial for the elderly. In this paper, a tinyML-based real-time system for detecting falls in elderly people was presented. Machine learning seems to be a perfect approach for the proposed system since it has the capability of learning from the available data and predicting the output observing various patterns and comparing them. Machine learning is perfect for this, but it usually needs a lot of power, which small embedded devices such as smartwatches do not have. That is where tiny ML comes in. This is a lighter version of machine learning that works well on small devices even with less processing power. Using tiny ML will make our proposed model strong and efficient compared to other similar existing models, with better performance and compromised processing.