Optimized Neural Network Based Location Prediction Along with Multiple Features in Communication Network

Abstract

By the advances in wireless communication networks and the exponential rise in the number of UEs, the data usage is increasing due to time consumption, necessitating lot denser deployment. Extra common handoffs result in superior latency and reduced throughput that have a deleterious impact on network and user acceptance. In this study, we have proposed a ResNet-based Convolutional Neural Network with Tasmanian Devil Optimization (TDO) algorithm for the prediction of location in the communication network. The ResNet-based CNN model with TDO algorithm is used for location prediction, taking into account of wireless quantification findings from service access point and neighbouring core network, and introducing a direction gradient descent to allow the prototype to recognize data on the direction of the UE motion. Comprehensive simulations proved that the proposed model, which is derived from various characteristics and communication flows, performed best result on the prediction of location.