Author Name: Kumar Dayanand, Udai Shankar, Mohit Kumar
Paper id: 26401
Abstract:
The possibility that Wireless Rechargeable Sensor Networks (WRSNs) can support longer operational lifetimes for wireless sensor networks (WSNs), as they provide mechanisms for both energy harvesting and replenishment, has recently received a considerable amount of interest. Therefore, a well-designed clustering technique is needed to maximize the usage of energy, reduce communication overheads and ensure that the entire network is stable. To achieve this goal, we have developed a hybrid clustering approach to effectively organize sensor nodes in a WRSNs through a combination of K-Means Clustering and the Gaussian Mixture Model (GMM). The initial step in our proposed hybrid clustering approach is K-Means clustering. K-Means clustering will rapidly group sensor nodes together based on their proximity to one another, thus ensuring an even distribution of workload across all clusters within the WRSN. Even though K-Means clustering is very efficient at grouping sensor nodes into clusters, it does not allow for overlapping clusters or probabilistic node placement within those clusters; therefore, we utilize GMM to enhance adaptability within dynamic network environments. We apply the Expectation-Maximization (EM) algorithm to iteratively optimize cluster membership probabilities to produce a better clustering scheme that is both accurate and adaptive. Performance metrics such as Silhouette Scores, energy consumption analysis and estimation of network lifespan will be utilized to compare the effectiveness of our proposed hybrid clustering approach with current techniques and demonstrate significant improvements in clustering efficiency and energy utilization. Overall, our proposed hybrid clustering strategy provides an efficient, flexible and scalable clustering solution for WRSNs, thereby providing improved resource management and extended lifetimes for sensor networks.