Machine Learning 5G Operations in HetNet

The wireless network is growing enormously which resulted in 5G, to improve the spectrum resources in the most feasible manner. The base station construction to cover the coverage holes is an infeasible approach, hence 5G has evolved into a multi-layer network based on various technological advances.

This multi-layer network offers an extensive range of wireless services. The deployment and operation of multi-tier wireless networks require a lot of planning, maintenance, and optimization. These multi-tier networks improve spectrum utilization by increasing the number of tiers, however, such a complex network design results in increased interference, backhaul, and QoS constraints.

The interference among devices is a growing concern in HetNets which is categorized as co-tier intra-cell interferences which require an accurate count of channel information. Not only this, but self-organization is also critical in a static network. However, it becomes a real challenge in dense environments where devices have conflicting objectives.

Apart from interference, the mobility of users affects the cooperation and strategic decision-making process. The mobile users in the network elevate the problem of resource management to a great extent, where traditional resource sharing and interference management solutions cannot be applied to achieve the QoS.

The mobility of the users can be used to predict network load which can help not only to manage the resources efficiently but also it can help to minimize interference. It is quite challenging to predict the model in a highly mobile heterogeneous environment and the key challenge here is to provide the most optimal solution with effective coordination.

This research aims to build a framework that leverages a machine learning solution to solve the interference issue in a resource constraint environment while considering the mobility factor.



Dr Muhammad Aman Sheikh
School of Engineering and Technology