Faulty Sensor Detection Approach for Wireless Sensor Networks Based on Artificial Intelligence and Information Theory

Wireless sensor network (WSN) is a main structure of wireless network made up of several linked sensor nodes. These nodes are sensors that are used to observe different environmental variables for the like humidity, wind speed, and temperature. WSNs are used in various applications, and the reliability of the collected data depends on the proper functioning of individual sensors. However, sensor failures and faults are inevitable, necessitating the development of effective faulty sensor detection approaches. This paper proposes an approach for detecting faulty sensors in WSNs based on a combination of logistic regression (LR), Machine Learning (ML) technique, and mutual information Variant (MIV). MIV method is the study of the correlation between the sensor and the neighbored sensor. While LR method presented for self-sensor detection. LR is mainly used for classification where the goal is to predict the sensor is faulted or not. The proposed algorithm demonstrates its ability to identify and classify faults accurately, thereby enhancing the overall performance and reliability of the network. The performance of the suggested algorithm was tested and applied on a real sensors data collected from a weather station at Benha, Egypt.