Support vector regressive dragonfly optimized shift invariant deep neural learning based handover for seamless data delivery in heterogeneous network

Abstract

In a Wireless Sensor Network (WSN), seamless mobility management can change the current mobile node’s location point to another network devoid of any link failure during communication. The seamless mobility system is very useful to detect the nearest base station over the wireless network without any distinct interference. In this paper, a novel technique called Support Vector Regressive Dragonfly Optimization based Shift Invariant Deep Neural Learning (SVRDO-SIDNL) is introduced for improving the seamless data transmission with minimum delay. The Shift Invariant Deep Neural Learning comprises of many layers to learn the series of input. For each layer, the different processes are carried out to accomplish the traffic optimized seamless data delivery. The input layer of the deep neural learning receives mobile nodes with coverage region and then is sent to the hidden layer. The mobile nodes' signal strength is analyzed by applying the support vector regression at the hidden layer. Then, the node with weak signal strength is identified and performs the handover. Through oppositional learned multi-objective dragonfly optimization technique, recognition of nearby attachment points with greater bandwidth availability is performed for the handover process. Then, the mobile node connection is altered from the existing attachment point to a new attachment point without losing connectivity. The simulation results reveal that the SVRDO-SIDNL technique offers a greater delivery rate, throughput with lesser packet loss at less delay.