Edge-AI in LoRa-based health monitoring: Fall detection system with fog computing and LSTM recurrent neural networks

Abstract

<p>Remote healthcare monitoring has exponentially<br />grown over the past decade together with the increasing penetration<br />of Internet of Things (IoT) platforms. IoT-based health<br />systems help to improve the quality of healthcare services<br />through real-time data acquisition and processing. However,<br />traditional IoT architectures have some limitations. For instance,<br />they cannot properly function in areas with poor or unstable<br />Internet. Low power wide area network (LPWAN) technologies,<br />including long-range communication protocols such as LoRa,<br />are a potential candidate to overcome the lacking network<br />infrastructure. Nevertheless, LPWANs have limited transmission<br />bandwidth not suitable for high data rate applications such as fall<br />detection systems or electrocardiography monitoring. Therefore,<br />data processing and compression are required at the edge of<br />the network. We propose a system architecture with integrated<br />artificial intelligence that combines Edge and Fog computing,<br />LPWAN technology, IoT and deep learning algorithms to perform<br />health monitoring tasks. In particular, we demonstrate the feasibility<br />and effectiveness of this architecture via a use case of fall<br />detection using recurrent neural networks. We have implemented<br />a fall detection system from the sensor node and Edge gateway to<br />cloud services and end-user applications. The system uses inertial<br />data as input and achieves an average precision of over 90% and<br />an average recall over 95% in fall detection.<br />Index Terms—IoT; Edge Computing; Healthcare Monitoring;<br />LoRa; LPWAN; RNN; LSTM; Fall Detection<br /></p>