Abstract
Wireless communication technologies play a key role in providing efficient and reliable Internet of Things (IoT) networks. Among them, Long Range (LoRa) technology and the LoRaWAN protocol are widely known for their ability to provide long-range communications with low power consumption and cost-effectiveness. One of the main challenging issues in deploying autonomous wireless networks is the ongoing need to optimize transmission parameters to minimize node energy consumption (EC) and maximize packet delivery ratio (PDR). This study introduces a novel transmission parameter selection algorithm such as Spreading Factor (SF) and Transmission Power (TP), leveraging machine learning (ML) methods such as XGBoost, GRU, and RBFN. The algorithm predicts the distance to a node based on the received RSSI, and subsequently forecasts EC and PDR, substantially enhancing network performance. The proposed approach demonstrates high prediction accuracy, achieving 99%, while reducing EC by 20.43% and increasing the PDR by 23.72% compared to the traditional adaptive data rate (ADR) algorithm