Autonomous Decentralized Transmission Timing Control in Wireless Sensor Network

In recent years, with the development of Internet-of-things (IoT) and machine-to-machine (M2M) communication, low power wide area (LPWA) networks have attracted attention. LPWA generally adopts a pure ALOHA protocol as the MAC layer access protocol to reduce the wireless device cost. If multiple wireless devices simultaneously transmit packets, a packet collision happens at a fusion center (FC) that collects wireless devices' information. To avoid such collision, the FC may control the wireless devices' transmission timings in a centralized manner. However, this requires control signal exchanges, resulting in increased communication overhead and increased battery consumption. In this paper, we propose a method to avoid packet collisions. The proposed method consists of two steps. Firstly, the transmission start timing of each device is autonomously set by reinforcement learning. Then, the transmission probability is determined based on the transmission start delay time. The computer simulation results show that the proposed method can improve the average packet delivery rate (PDR) by 42% compared to the conventional ALOHA protocol.