Practical evaluation of carrier sensing for a LoRa wildlife monitoring network

We consider the technique of carrier sensing for application in a LoRa mesh network aimed at wildlife monitoring. A key challenge in this application is to limit collisions in order to increase the channel capacity. Since CSMA is very rarely applied in LoRa-based networks, our goal is to determine its practical viability. We evaluate the LoRa Channel Activity Detection (CAD) mechanism under laboratory and field conditions. Our results show that both preamble and payload symbols are detectable even at distances exceeding 4 km. Detecting LoRa preamble symbols had a SNR advantage of between 1 and 2 dB over payload symbols. Furthermore, we find that by taking at least 8 consecutive CAD measurements during payload frame time, a clear channel assessment (CCA) comparable to the LoRa frame reception rate can be achieved between two nodes.