A Time-Frequency Geometric Detection Approach for LoRa Chirp Signals

Long Range (LoRa) defines a particular chirp spread spectrum modulation employed for signaling in the LoRaWAN network. LoRa is recognized as one of the most effective technologies for the Internet of Things communication scenario, providing high reliability, wide coverage and low energy consumption for devices. The achievement of such performance is tied to the capability of LoRa receivers to accurately detect the signal even at low signal-to-noise ratio conditions. Dealing with signal demodulation, coherent and non-coherent detection represent the most widespread approaches. In this contribution, we present a new detection mechanism that exploits the geometrical properties of linear chirp to perform symbol demodulation, without requiring the energy consuming operations related to coherent and non-coherent detection. Furthermore, we demonstrate how the proposed approach can be fruitfully exploited not only for symbol demodulation, but also for LoRa packet synchronization and spreading factor recognition. By showing some numerical results, we compare the reliability level of the presented scheme through bit error rate with respect to the one offered by conventional detection methods.