| Autor: |
Taras Maksymyuk, Juraj Gazda, Madhusanka Liyanage, Longzhe Han, Bohdan Shubyn, Bohdan Strykhaliuk, Oleh Yaremko, Minho Jo, Mischa Dohler |
| Rok vydání: |
2021 |
| Předmět: |
|
| Zdroj: |
Lecture Notes in Electrical Engineering ISBN: 9783030924331 |
| DOI: |
10.1007/978-3-030-92435-5_19 |
| Popis: |
A smart architectural design in 5G with flexibility for various deployment scenarios and service requirements has enabled different business models for mobile network operators in both nationwide and local scales. Future 6G networks will feature even more flexible mobile network deployment driven by spectrum and infrastructure sharing among the operators. In this chapter, we propose a new multi-layer framework for 6G with decoupled operators and infrastructure planes. The proposed framework provides a flexibility of network configuration for multiple operators in condition of open spectrum and infrastructure market by using a multi-dimensional matrix representation of the data flows. In particular, the proposed model supports the dynamic switching of the operator and multi-operator service provision for the end users. As a case study, we have developed an AI-based workflow for the dynamic spectrum allocation among multiple mobile network operators. The key advantage of the proposed workflow is that it can be adjusted to the different combinations of the data flows and thus can be suitable for the spectrum allocation among multiple operators. The intelligent capabilities of the proposed workflow are provided by the deep recurrent neural network based on the long short-term memory architecture. The developed model has been trained over the custom dataset with realistic user mobility in urban area. Simulations results show that the proposed intelligent model provides a stable service quality for end users regardless of the serving operators and outperforms the static and semi-intelligent models. |
| Databáze: |
OpenAIRE |
| Externí odkaz: |
|
