Quadratic ensemble weighted emphasis boosting based energy and bandwidth efficient routing in Underwater Sensor Network

Abstract

Underwater Sensor Network (UWSN) is a network that comprises a large number of independent underwater sensor nodes to perform monitoring tasks over a given area. UWSN minimized propagation delay, bandwidth, and packet loss. However, the implementation of efficient communication is a significant problem at UWSN. Therefore, Energy and Bandwidth-aware Quadratic Ensemble Weighted Emphasis Boosting Classification (EB-QEWEBC) method for performing energy-efficient routing in UWSN is proposed. Initially, different numbers of underwater sensor nodes are considered as input. Next, the bandwidth and energy consumption of every underwater sensor node is measured. After that, classification between underwater sensor nodes is made by considering energy and bandwidth as factors using Regularized Quadratic Classifier (i.e., weak classifier) for performing routing with minimum delay. Followed by, Weighted Emphasis Boosting is utilized to ensemble weak learners to form strong learners for improving data routing performance results with the biconvex combination. Finally, after classifying the node, data packets are sent to higher energy and bandwidth-efficient underwater sensor nodes. The classification process is carried out at every underwater sensor node for transmitting data packets to the sink node with minimum delay. This method determines the energy-efficient data communication through classification and boosting to reduce the misclassification rate. Experimental results EB-QEWEBC shows a minimization of 14%, 21%, 26%, and 54% in terms of bandwidth, energy consumption, end-to-end delay, and misclassification rate as compared to state-of-art-methods respectively.