| Autor: |
Silva EF; Department of Computer Science, Federal University of Juiz de Fora (UFJF), Juiz de Fora 36036-330, MG, Brazil., Figueiredo LM; Graduate Program in Computer Science-UFJF, Juiz de Fora 36036-330, MG, Brazil., de Oliveira LA; Department of Computer Science, Federal University of Juiz de Fora (UFJF), Juiz de Fora 36036-330, MG, Brazil., Chaves LJ; Department of Computer Science, Federal University of Juiz de Fora (UFJF), Juiz de Fora 36036-330, MG, Brazil., de Oliveira AL; Department of Computer Science, Federal University of Juiz de Fora (UFJF), Juiz de Fora 36036-330, MG, Brazil., Rosário D; Faculty of Computer Engineering (ENGCOMP), Federal University of Pará (UFPA), Belém 66075-110, PA, Brazil., Cerqueira E; Faculty of Computer Engineering (ENGCOMP), Federal University of Pará (UFPA), Belém 66075-110, PA, Brazil. |
| Abstrakt: |
Sub-GHz communication provides long-range coverage with low power consumption and reduced deployment cost. LoRa (Long-Range) has emerged, among existing LPWAN (Low Power Wide Area Networks) technologies, as a promising physical layer alternative to provide ubiquitous connectivity to outdoor IoT devices. LoRa modulation technology supports adapting transmissions based on parameters such as carrier frequency, channel bandwidth, spreading factor, and code rate. In this paper, we propose SlidingChange, a novel cognitive mechanism to support the dynamic analysis and adjustment of LoRa network performance parameters. The proposed mechanism uses a sliding window to smooth out short-term variations and reduce unnecessary network re-configurations. To validate our proposal, we conducted an experimental study to evaluate the performance concerning the Signal-to-Noise Ratio (SNR) parameter of our SlidingChange against InstantChange, an intuitive mechanism that considers immediate performance measurements (parameters) for re-configuring the network. The SlidingChange is compared with LR-ADR too, a state-of-the-art-related technique based on simple linear regression. The experimental results obtained from a testbed scenario demonstrated that the InstanChange mechanism improved the SNR by 4.6%. When using the SlidingChange mechanism, the SNR was around 37%, while the network reconfiguration rate was reduced by approximately 16%. |
