Low Power Consumption Online Testing Technique for Wearable Wireless Body Area Networks

Internet of Things (IoT) is a growing field of research and in real-life applications. Cyber-physical Systems are aimed to monitor physical variables, processing data, applying artificial intelligent algorithms, and responding in real time for further actions. For example, on-the-edge medical monitoring of chronic illnesses or supervising of an industrial process can help to early detect small anomalies or misfunctions, preventing catastrophic events or assuring minimum quality thresholds. These systems require continuous monitoring, data collection, signal processing and analysis, which demand smart sensors, efficient wireless communications, and powerful processing modules. This is especially critical in wearable body area networks (BAN) where size, weight and power consumption are key elements in the system's usability and attractiveness. Dependability becomes a big concern in the design and deployment of wearable devices that can be considered unsupervised wireless nodes, with high complexity in architecture and functionality. This work proposes the integration of a software online testing procedure for monitoring intra-node faults appearance, detecting the presence of errors, reporting their effects, and performing a graceful degradation when possible, maintaining a low power consumption. The online testing proposal has been implemented and validated in a real system, Bindi from the UC3M4Safety team. The effect on the power consumption has been measured showing a very low effect on the overall power budget.