An SDR Implementation of Frequency Domain Demodulation of LoRaWAN Messages with Temporal Correlation

This paper presents a practical LoRaWAN system serving applications that generate time-correlated data, among other applications. For instance, a smart meter reading at a given time can be assumed to be modeled as a random increment from the previous reading. The paper reports a demodulation technique that can be deployed along with existing techniques to detect transmissions from such applications, through a prediction and frequency matching based packet decoding algorithm. We also demonstrate the working of our approach on a software defined radio platform. Practical aspects like energy detection and carrier frequency offsets for LoRa signals are addressed. Limitations of the proposed scheme are provided.