Modified Deep Learning Model in Proactive Decision-Making for Handover Management in 5G

Abstract

With the fast expansion of mobile devices and internet traffic, it is becoming critical to deliver dependable and robust services. HetNets and large networks are highlighted as probable solutions to the nearing capacity obstructions; however, they also present substantial challenges in terms of handover (HO) management. In cellular telecommunications, HO describes the procedure of moving an active call or data link from one base station (BS) to another. Whenever a mobile phone switches to another cell while a conversation is in progress, the MSC (mobile switching center) shifts the call to an alternate channel assigned to the new BS. The major objective of this work is to assist in how the HCP includes the functions of the 5G network, in which a modified deep learning architecture is introduced for predicting the NDR (network download rate) efficiently. In particular, a modified DNN architecture is introduced for this purpose. As a result, the proposed model attained a lower HO delay of 10.207 ms at a speed of 60 m/s, which surpasses the results of established techniques. From the analysis, it is proven that the proposed work efficiently increases the performance of the network without any interruption during transitions among cells.