Lora Network Planning Based On Improved Isodata Algorithm

LoRa is a well-known Low-Power Wide-Area Network (LPWAN) specification that often uses a star topology to directly connect IoT devices to the gateway, which sends data to the network server. In practical deployment, determination of the number of gateways and their locations is the key problem in LoRa network planning. In this paper, we model LoRa network planning as a constrained clustering problem and propose an improved ISODATA algorithm to solve it. Simulation results show that, the packet reception rate and the average throughput can be improved by 23.30% and 23.51% compared with the basic ISODATA algorithm.