Applying reinforcement learning in slotted LoRaWAN: From concept to implementation
Zorbas D. Kasenov S. Salimzhanova K. Gaziz D. Ismailov T. Baimukhanov B.
1 October 2025Elsevier B.V.
Computer Communications
2025#242
As Low Power Wide Area Networks (LPWANs) are increasingly adopted for Internet of Things (IoT) applications, they face significant challenges related to interference and scalability, which can lead to high collision rates and reduced network throughput. This paper presents a novel approach to enhancing the performance of LoRaWAN, one of the dominant LPWAN protocols, by leveraging Reinforcement Learning (RL). The proposed solution introduces a synchronization framework designed to operate under LoRaWAN principles, coupled with a low-cost, on-device RL mechanism that autonomously mitigates collisions. Through extensive simulations and real-world experiments, the effectiveness of the RL approach is demonstrated, showing an over 30% improvement in terms of packet delivery ratio (PDR) compared to traditional multiple access methods such as Pure-Aloha, Slotted-Aloha, and Carrier Sense Multiple Access (CSMA). Additionally, open-source implementations for both simulation and experimental validation are provided, ensuring reproducibility and facilitating further research in this domain.
Internet of Things , LoRaWAN , Reinforcement learning
Text of the article Перейти на текст статьи
Nazarbayev University, School of Engineering & Digital Sciences, Astana, Kazakhstan
Nazarbayev University
10 лет помогаем публиковать статьи Международный издатель
Книга Публикация научной статьи Волощук 2026 Book Publication of a scientific article 2026