LoRaWAN Gateway Placement in Smart Agriculture: An Analysis of Clustering Algorithms and Performance Metrics

The use of Wireless Sensor Networks (WSN) in smart agriculture has emerged recent years. LoRaWAN (Long Range Wide Area Networks) is widely recognized as one the most suitable technologies for this application, due to its capacity transmit data over long distances while consuming little energy. Determining number and location gateways (GWs) a production setting challenging tasks planning building type network. Various solutions gateway placement problem have been proposed literature, utilizing clustering algorithms; however, few works compared performance various strategies. Considering all these facts, paper proposes strategy localization GWs, cover vast agricultural region. Four algorithms were used deploy network GWs: K-Means three versions: Minibatch K-Means; Bisecting Fuzzy c-Means (FCM). As metrics, uplink delivery rate (ULDR) energy consumption used, provide subsidies designer client, with which choose best setup. A stochastic model was evaluate power consumption. Simulations performed, considering two scenarios: Scenario 1 lower-medium concurrence, 2 higher-medium concurrence. simulations showed that more than GWs did not lead significant improvements ULDR consumption, whereas, 2, suggested between 11 15. results 1, FCM algorithm superior alternatives, regarding mean respect maximum In relation caused produced lowest