A Weighted Moving Average adaptive technique for LoRa-based networks

Abstract

The rise of Internet of Things (IoT) with a wide range of innovative applications such as autonomous vehicles, smart cities, and smart homes demands reliance on wireless communication to collect and transmit a massive amount of data. Recent advances in wireless technology, such as Low-Power Wide Area Networks (LPWAN) and Sub-GHz communication protocols (e.g., LoRaWAN) have been demonstrated to be effective in supporting long-range (LoRa) connectivity between IoT devices with low power consumption and reduced deployment costs. Although LoRaWAN signal modulation enables adapting transmissions based on parameters (e.g., Signal to Noise Ratio - SNR), as the adoption of the communication technology in a given frequency spectrum increases, the transmission medium dispute and interference also increase. The Cognitive Radio (CR) supports the dynamic adjustment of LoRa network parameters to mitigate interference and network congestion. In this master’s thesis, a Weighted Moving Average (WMA) adaptation technique is proposed to reduce network congestion by improving SNR, and minimizing packet loss in LoRa-based networks of IoT devices. A controlled experiment was conducted in a university testbed network environment to evaluate the feasibility of the proposed WMA modulation technique. The results demonstrated that the proposed WMA technique reduced 5.65% the network reconfiguration time, and improved the SNR at 39.09% compared to Sliding Change and LR-ADR (Long-Range Adaptive Data Rate) state-of-the-art techniques.