Capacity Planning of LoRa Networks With Joint Noise-Limited and Interference-Limited Coverage Considerations

LoRa networks are emerging low power wide area networks (LP-WAN) that are configured as star-of-star networks. Within each underlying star network, potentially thousands of LoRa end-devices utilize the patented chirp spread spectrum-based LoRa modulation to communicate directly with a LoRa gateway over long distances at the scale of kilometers. Typically, each LoRa gateway supports six quasi-orthogonal logical star networks that correspond to the available LoRa modulation spreading factors. The aggregate capacity of a LoRa gateway is evaluated as the number of end-devices that could be supported within the corresponding cell area by all six available spreading factor networks while fulfilling some QoS coverage requirement. Successful coverage of LoRa signals could be mainly attributed to noise-limited and interference-limited considerations. In the LoRa literature, these two factors have been treated in an isolated manner, where independent SNR and SIR thresholds have been set to signify coverage. However, in reality, the SNR threshold for successful coverage depicts a decreasing function of the SIR experienced by the end-device of interest. In this paper, the capacity of LoRa networks is evaluated with joint noise and interference considerations. The presented framework for capacity evaluation has the advantage that it derives the cumulative capacity distribution that could be supported from the cell-edge towards the cell-center of an individual LoRa gateway. It is then demonstrated that the capacity distribution curves could be employed in the LoRa network planning of an example IoT home security application.