A Unified Network Coding Framework for LR-FHSS Headers and Payloads

Long-Range Frequency-Hopping Spread Spectrum (LR-FHSS) improves the upload capacity of LoRaWAN networks by fragmenting the payload and transmitting multiple header copies. This method, along with frequency hopping, helps towards mitigating interference. Previous works have identified headers as the primary bottleneck in LR-FHSS networks, and then proposed techniques to enhance header protection. However, that assumption is not universally valid and depends on factors such as payload length, number of end-devices, and channel conditions. In scenarios with small payloads, vulnerable links or high network density, the payload's error probability can become the dominant factor. This paper first identifies this gap, and then proposes a novel scheme, Full Network Coding (LR-FHSS-FNC), which extends the header network coding concept recently proposed in the literature also to the payload fragments. Our results show that LR-FHSS-FNC provides considerable performance gains precisely in these payload-dominant scenarios, further enhancing network reliability and scalability without requiring additional spectral or power resources.