Fuzzy Controlled Wavelet-Based Edge Computing Method for Energy-Harvesting IoT Sensors

Abstract

The study presents a novel edge computing (EC) method based on a discrete wavelet transform (DWT) and fuzzy logic controller suitable for application with energy harvesting IoT sensors. The authors propose a new solution to address information latency in an IoT device when compressed data with high information density are transmitted to the cloud with high priority or detailed information is added to the cloud when the energy state in the IoT device is sufficient. The solution potentially delivers a completely lossless scenario for low-power sensors, a significant benefit that state-of-the-art methods do not provide. The article describes the hardware model for an IoT device, input and predicted energy data, and a methodology for designing the parameters of DWT and fuzzy logic controllers. The results of the study indicate that the proposed EC method achieved full data transmission in contrast to the reference solution which had the worst case parameters of maximum outage and penalties caused by delayed data. The average delay in uploading approximate data was 0.51 days with the proposed fuzzy controller EC method compared to reference methods, which have an average delay of at least 0.91 days. The results also highlighted the importance of the trade-off between information latency and reliable functionality. The results are discussed in terms of an innovative approach which features an IoT sensor that maximizes its own energy consumption according to the data measured from specific parameters.