Adaptive FCM‐Based Cluster Head Selection and Secured Data Transmission by Deep SARSA on Wireless Sensor Network

Abstract

In modern times, wireless sensor networks (WSNs) are necessary for remote system surveillance and information exchange. Despite the advantages of WSNs, node lifetime and data loss limit their use. Several researchers use energy‐efficient protocols to solve the aforementioned issues in WSNs. However, the majority of authors focused exclusively on the energy‐efficient model for solving the issues. Yet, this results in high energy consumption and repetitive processes. So, this research developed a cluster head (CH) selection method using the fuzzy model for effective data transmission. The main contribution of the proposed work is to perform secure data transmission within the WSN by reducing and balancing energy consumption. Initially, the necessary network attributes are gathered from the available resources. The adaptive fuzzy C‐means clustering (AFCMC) concept is employed in clustering the WSN. In the cluster‐based WSN, CH is optimally chosen with the support of the improved black widow optimization algorithm (IBWOA). Here, the parameter in AFCMC is optimally tuned by the same IBWOA. The tuned parameters include the fuzziness parameter (2–20), epsilon (1–10), and the number of iterations (10–100). Subsequently, the data transmission is conducted with the support of the deep state–action–reward–selection algorithm (Deep SARSA). In this, the suggested Deep SARSA helped to perform the data transmission securely through the shortest path and improved the detection of malicious nodes. The newly designed approach is compared with the solutions in the literature. The established IBWOA‐AFCMC models in terms of residual energy are 16.66%, 21.73%, 19.14%, and 55.55% more effective than KMC, hierarchical clustering, biclustering, and FCM in the 100th number of nodes. This dual‐layer approach significantly enhances network longevity and data integrity in dynamic WSN environments.