Countering DoS Threats in WSN: Inventive Machine Learning Strategies with a Minimalistic Twist
Senthilnathan Chidambaranathan, Vanitha, C.S. Sasireka, Naduvathezhath Nessariose Jose, Kanduri Madhusudhana Rao, ThIbetombi Devi · 2024
In the era of expanding wireless sensor network (WSN) applications, safeguarding these networks against a spectrum of threats, including Denial-of-Service (DoS) attacks, is paramount. The innate resource limitations of sensor nodes often render conventional security measures ineffective. This study introduces a pioneering approach integrating minimalist machine learning techniques to combat DoS attacks within WSNs. Our methodology adeptly leverages machine learning capabilities while strategically curbing computational and resource demands. By distilling essential features from network traffic patterns, our model adeptly identifies aberrations signaling DoS attacks, all while preserving the efficiency of resource-limited sensor nodes. Its minimalist framework seamlessly integrates with existing WSN architectures, negating the need for substantial hardware upgrades and maintaining network performance. Results underscore the superior detection accuracy and efficiency of our minimalist machine learning strategy compared to traditional methods. Moreover, its scalability and adaptability render it ideal for deployment across varied WSN setups, fortifying them against DoS threats while safeguarding core network functionalities.By mitigating DoS attack impacts without burdening computational resources, our approach marks a significant stride towards bolstering the security and dependability of wireless sensor networks amid evolving threat landscapes.