Analysis of Dynamic ADR Algorithms for LoRa Networks in IoT Environments

Due to a large number of end-devices and channel noise in the Internet of Things (IoT) environments, it may reduce the packet delivery rate and increase energy consumption when using the traditional adaptive data rate (ADR) algorithm of LoRa networks. Considering this problem, a new dynamic ADR algorithm is proposed for the IoT environments that perform joint weighting processing on the dynamic link Margin and the Ordered weighted average of signal-to-noise ratio (SNR), namely ADR-MO, which can dynamically configure spreading factor and transmission power to adapt to different channels. Then, an RL-ADR algorithm using reinforcement learning is proposed. The results show that the ADR-MO algorithm can quickly adapt to sensitive changes in the channel state and has good performance with the increase of the number of end-devices and the effect of shadow fading. It also has a higher packet delivery ratio and lower average energy consumption in the 470-510 MHz frequency band. Besides, the RL-ADR algorithm can respond to state changes faster and has advantages with a few end-devices.