Detecting Inaccurate Sensors on a Large-Scale Sensor Network Using Centralized and Localized Graph Neural Networks

This article conducts an empirical study on detecting faulty sensors in a large-scale sensor network containing approximately
$10\,000$
sensors distributed over
$36\,000 \text {km}^{2}$
. First, we discuss the practical challenge of this task. We compare rule-based models, traditional machine learning models, deep learning models without graph neural networks (GNNs), and deep learning models with GNNs. The experimental results show that GNNs identify more problematic sensors in fewer trials than rule-based models and other machine learning and deep learning models. In addition to training the models in a central server, we also show that localized versions of the deep learning models with GNNs yield predictive power comparable to centralized training. Consequently, each sensor may perform a local inspection to identify its health status and only send reminder signals to a centralized server if it is self-diagnosed as a faulty sensor.