DDoS Attack Detection in WSN Using Modified BGRU With MFO Model

Abstract

Significant challenges in the areas of energy and security persist for wireless sensor networks (WSNs). Avoiding denial-of-service assaults is a priority for safeguarding WSN networks. As open field encryption becomes the norm, conventional packet deep scan systems can no longer use open field review in layer security packets. To the existing literature evaluating the effect of deep learning algorithms on WSN lifespan, this study contributes the auto-encoder (AE) and then the bidirectional gated recurrent unit (BGRU). The learning rate of the BGRU is also chosen using the moth flame optimization technique. Learning is just one of the approaches that have emerged in response to the pressing need to distinguish between legitimate and criminal users. This chapter also demonstrated that for numerical statistical data, the sweet spot is reached when the number of records in the dataset is between three thousand and six thousand and when the percentage of overlap across categories is not less than fifty percent.