DBKNN and Radial-ANFIS Model for Energy Efficient Wireless Sensor Network

Abstract

In a Wireless Sensor Network (WSN), packet transmission and sensing functions are the most energy consumption factors. When it comes to wireless communication, unnecessary sensing could increase data communication which in turn needs the biggest amount of energy for communication. Thus, to minimize Energy Consumption (EC), various research has been conducted. Still, they don’t provide efficient EC. To solve this issue, a novel Davies-Bouldin K Nearest Neighbor (DBKNN) and Radial-Adaptive Neuro Fuzzy Interference System (Radial-ANFIS) model is proposed for energy-efficient WSNs. Initially, partitioning is performed on deployed nodes. Then, node features are extracted. By employing Laplacian Cubic Cheetah Optimizer (LCCO), CHs are selected using features; then, nodes are clustered using DBKNN. Next, weights are assigned to each clustered node. By employing Adaptive Intensed Golden Tortoise Beetle Optimizer (AI-GTBO), relay nodes are selected using weight value. Paths are created betwixt the node and Base Station (BS) utilizing relay nodes; then, the shortest path is computed. Nodes sense the data once the network is designed, and the correlation between each data is calculated. For redundant node detection, the analytic output is given to Radial-ANFIS. Lastly, redundant nodes are kept to sleep, and the remaining node data are utilized for transmission. When data is transmitted to BS, its first dimension is reduced using Newton Raphson Iterative Principal Component Analysis (NRI-PCA). According to outcomes, the proposed model reduces EC and maximizes the overall throughput and network lifetime compared to other methods. The EC rate is also reduced to 1364 J using the proposed method.

Keywords

• AI-GTBO
• Davies-Bouldin K Nearest Neighbor
• Laplacian Cubic Cheetah Optimizer
• NRI-PCA
• Radial-ANFIS