AI-Enhanced Intrusion Detection and Cluster Head Selection for Quality of Service (QoS) Optimization in Wireless Sensor Networks

Abstract

The widespread relevance of wireless sensor networks (WSN) has made them an active area of research. Quality of service (QoS) considerations in WSN design center around limiting power consumption, increasing network longevity, and bolstering safety. The dynamic nature of a Mobile Ad hoc Network (MANET) makes it vulnerable to issues with both power and safety. Optimization of energy is a difficult problem to solve for most of the already available methods, yet routing protocols do it efficiently. Many issues, such as path maintenance, power consumption, security, reliability, and sudden changes in connection characteristics, arise because of the mobility of nodes in MANETs. Current clustering methods suffer from unreliable communication that wastes the power of nodes due to the fact that the interaction lacking from the model by the possibility of dropped packets. Therefore, improving system QoS necessitates employing energy-efficient communication methods. The network's dependability may be greatly enhanced by deploying the AI – Enhanced Multicast Routing Technology [AI-EMRT]. Multicasting involves sending data packets from a source node to multiple receiving nodes at once. Multicasting is a cost-effective way to send data. QoS criteria including energy, longevity, and security are developed in this research with a novel code for safe unequal clustering using an intrusion detection system. The suggested model begins by selecting the Tentative Cluster Heads (TCHs) via a Dynamic Cognitive Fuzzy Driven Clustering Approach [DCFDCA] with the help of residual energy, BS distance, and neighbor distance are all possible input variables. Finally, a successful intrusion detection system based on a dense web of convictions is implemented on the Cluster Heads (CH) to detect the invasion of privacy in a cluster-based WSN, ensuring its security. The higher energy efficiency, network lifetime, packet delivery percentage, mean delay, and intrusion detection performance of the proposed approach rate has been shown by an extensive set of experiments.