Application of Neural Network in Handover Predictions and Resource Allocation in Long Term Evolution

Abstract

This work took handover enhancement into consideration. Better handover performance is attained with the aid of two Artificial Intelligence (AI) entities. Less frequent handovers occur when the load is evenly distributed across the SeNodeBs. The suggested load balancer was built on an artificial neural network clustering model that uses a self-organizing map as a hidden layer. It was trained to predict network conditions, minimize handovers—especially for UEs at the cell edge—by carrying out only those that were absolutely necessary, and steer clear of handovers to the Macro cell for downlink directions.Hold revolving in the handover orbit, another way to keep and make use of network assets was by predicting the handovers before they arise, and allocate the desired information inside the target SeNodeB, The predictor entity within the proposed gadget architecture combined the features of Radial basis characteristic Neural community and neural community time collection tool to create and replace prediction list from the system’s amassed data and learnt to predict the following SeNodeB to companion with. The prediction entity simulated the usage of MATLAB, and the effects showed that the machine was capable of supply as much as 92% accurate predictions for handovers which brought about universal throughput improvement of 75%.