Intelligent Configuration of PHY-Layer Parameters to Reduce Energy Consumption in LoRa

Communications over long distances and strong resilience to interference are vital aspects of LoRa. LoRa adjusts the modulation to allow higher data transmission rates, depending on the reception sensitivity threshold and the communication distance. The spreading factor and the transmission power, in turn, are directly related to energy consumption, influencing network performance. This paper proposes the use of supervised learning techniques to conFigure the spreading factor and the transmission power simultaneously. This approach differs from the literature as it configures two parameters instead of just one, the spreading factor. Different learning techniques are evaluated through simulations using a LoRa network. Our experiments compare the performance of our proposal with the traditional LoRaWAN and the state-of-the-art on intelligent configuration using only the spreading factor. The obtained results show that our proposal successfully reduces the energy consumption without affecting the packet delivery ratio.