Downlink spectrum allocation in 5G HetNets

Fifth generation mobile systems (5G) target an Average Area Spectral Efficiency (AASE) over hundred Gbps/km²/user for future mobile systems with an Energy Dissipation (ED) per unit area similar to the current ED levels. Heterogeneous networks (HetNets) with high density of deployed small cells are currently adopted to aid in achieving the target ED and AASE by 5G. Limited spectrum availability requires efforts to manage the spectrum utilization in such dense deployments. Development of new network architectures and Radio Resources Management (RRM) schemes is important to address such challenges. The objective of this work is to propose a new architecture that consists of a Decision Support System (DSS) and a data collection system to dynamically manage and control the spectrum allocation process. The DSS generates spectrum allocation patterns using non-parametric estimation and statistical analysis for the collected data. A new RRM model using a Plan, Do, Control and Act (PDCA) cycle is proposed as a new self optimization module in the self organizing network framework. The PDCA model utilizes the new architecture and the allocation patterns to dynamically predict future spectrum allocation. Results show improvement in the AASE achieved using the PDCA model compared to conventional spectrum allocation.