An MDP model-based initial strategy prediction method for LoRaWAN

As one of the technologies in the wide-area network category, LoRaWAN provides a wireless network with a large capacity of end devices (ED) in long-range. With a pure Aloha protocol implemented into its MAC layer, LoRaWAN can reduce its power consumption. Besides, some orthogonal transmission parameters give LoRaWAN capability to avoid collision and packet loss. Thus, allocating transmission parameters to increase the network performance becomes a challenging issue for LoRaWAN. Some dynamic Spreading Factor (SF) allocation strategies are studied in this paper. A distributed Markov Decision Process (MDP) model is constructed for the uplink transmission of the class-A device in LoRaWAN. The model is also solved and implemented to the algorithms for the initial strategy prediction. Analytical results show that the MDP model increases the performance of the studied algorithms on the transmission of the packet.