Dataset of "Feasibility of Commercial Microwave Links as a Network of Distributed Temperature Sensors"

This paper investigates the feasibility of commercial microwave links (CMLs) as distributed sensors for temperature monitoring. Our analysis reveals that CML temperature readings strongly correlate with ambient temperatures recorded by a nearby weather station. However, several factors, including heat produced by device operation, direct sunlight, convection, and resets introduce offsets and temperature fluctuations that must be taken into account to ensure reliability and output accuracy. Using a feature selection method, we rank the CML data according to their significance for temperature prediction and demonstrate three machine learning models to predict the current ambient temperature. Although CMLs offer a cost-effective and high-density monitoring solution without additional hardware costs, their readings require offset corrections to estimate environmental temperature. In this pilot study, we demonstrate various advantages and pitfalls of this approach and discuss future steps for successful applications.