Intelligent Fault Detection in IoT-Based WSNs for Precision Agriculture Using Machine Learning

Abstract

Ensuring the reliability of data in large-scale Things IoT-based Wireless Sensor Networks deployed over extensive areas presents a significant challenge, particularly concerning the detection and classification of faulty sensor data. This paper investigates the application of machine learning models for intelligent fault detection in such environments. Based on a database encompassing diverse fault types and varying fault rates, several prominent learning algorithms, including Support Vector Machine, Random Forest, Naïve Bayes, and Decision Tree, were trained and evaluated on this faulty dataset. The performance of these models in accurately identifying faulty data under different fault scenarios and densities was rigorously assessed by comparing the detection accuracy and F1-score of these algorithms. This study aims to determine the most robust and effective ML model capable of maintaining high fault detection accuracy across varying fault conditions, thereby enhancing the overall data reliability for critical large-scale IoT applications.