MeanC-IQR: Enhanced Fault Detection in WSNs for Smart Farming

Abstract

In smart farming, reliable data from Wireless Sensor Networks (WSNs) is critical for informed decision-making. This paper proposes the MeanC-IQR approach, a novel fault detection (FD) technique that enhances the accuracy and robustness of data collection in WSNs. The method leverages the strengths of both centralized mean (MeanC) analysis and Interquartile Range (IQR) analysis in a dual-layer approach. This synergy effectively identifies various fault types while minimizing false positives. Additionally, the MeanC-IQR approach maintains low computational overhead, making it suitable for resource-constrained WSN deployments in smart farming applications. Extensive simulations and real-world evaluations demonstrate the effectiveness of MeanC-IQR, showcasing its superior performance in detecting sensor faults compared to existing methods.