Revised Gateway Selection for LoRa Radio Networks

Abstract

As many research papers show, one of the problems of a LoRa network is its limit regarding the scalability. However, these papers also indicate that it is possible to achieve the scalability for them by dynamically selecting transmission parameters and/or by employing multiple gateways. In this paper, we build upon the latter solution and show that although data extraction rate in such networks is quite good, it suffers hugely from data duplication on the communication path between the gateways and network server. In remote areas, the gateways are usually connected to a network server by cellular networks, which yields additional transmission costs. In LoRaWAN network topology, the gateways aren’t connected to each other and therefore cannot actively filter the traffic prior to sending it further downstream. Thus, we propose a randomized algorithm allowing to reduce duplication of the packets sent to the network server without communicating with other gateways.