LoRa Device 의 이동성을 고려한 GRU 기반 적응형 데이터 전송 속도 제어 알고리즘

Abstract

Long Range Wide Area Network (LoRaWAN), a representative technology of Low Power Wide Area Network (LPWAN), is a networking technology for the long-range Internet of Things. It features low power, a long distance, and low speed, and each device in LoRaWAN supports adaptive data rate (ADR) technology to save transmission power. For implementing ADR in LoRaWAN, a network server uses algorithms to reduce power consumption by minimizing the time LoRa Radio packets stay in the air and sends a medium access control (MAC) command frame to increase the data rate or adjust transmission power by referring to signal-to-noise ratio (SNR) values. However, because the existing ADR algorithm for the network server sets the appropriate data rate and transmission power based on the maximum SNR value of the recent 20 packets, it does not support mobile devices because the SNR values of the mobile devices vary over time depending on their speeds. This paper introduces an improved ADR, GRU-ADR, that infers the future SNR values using the deep-learning gated recurrent unit (GRU) model to set appropriate data rates and transmission power using the ADR function even in mobile devices. The simulation study based on the OMNeT++ simulator and the Framework for LoRa (FLoRa) shows that GRU-ADR outperforms the existing ADR in packet delivery ratio (PDR) and energy consumption.