The research focuses on hierarchical clustering-based intrusion detection using artificial neural networks (ANNs) for secure data transmission in wireless sensor networks (WSNs). WSNs consist of numerous tiny sensor nodes deployed to monitor environmental phenomena. These nodes face significant challenges, including restricted energy, memory space, communication range, and limited capacity for managing energy, storing, transmitting, and processing data. To address these limitations, a machine learning-based approach is proposed to detect intrusions and efficiently utilize energy by properly selecting cluster heads using a secure clustering protocol. The proposed method was implemented and tested using MATLAB software, employing the NSLKDD and UNSW-NB15 datasets for intrusion detection. The results demonstrated promising outcomes in detecting intruders and enhancing network efficiency, achieving a 92% packet delivery ratio (PDR) and 1.82 Mbps throughput. The study concludes that while WSNs are gaining popularity due to their simplicity, flexibility, and scalability, innovative solutions are necessary for efficient energy management and security. Future research should focus on advanced machine learning models, energy harvesting techniques, scalable protocols, real-time data processing, and integration with IoT platforms for broader applications and enhanced functionality.
Read full abstract