Deep Learning based Adaptive Handover Optimization for Ultra-Dense 5G Mobile Networks

Abstract

An overview is devoted to the automation of fifth- generation mobile communications based on the use of artificial intelligence. We suggest using GRU recurrent neural networks, as they provide a rapid response to changes in the environment, which is often case in the wireless networks. It is also proposed to use a three-tier model to integrate artificial intelligence into the mobile network, which will effectively increase the amount of useful information transmitted on the channel and reduce service information, due to the fact that on each device containing a block with a neural network all personal and sensitive information will be processed locally, and only the results of neural networks will be sent to the main Knowledge Base server, which will only be suitable for further processing by the neural m network. This approach will significantly reduce the amount of service traffic that will be transmitted through the communication channels.