Tutorial 42022-06-06T10:11:04+00:00

Deep Learning and Reinforcement Learning for Future Wireless Network Optimization

Tuesday, 7 June 2022, 14:00-15:30/16:00-17:30, Room A228
Speaker:
  • Haijun Zhang (University of Science and Technology Beijing, CN)
  • Yansha Deng (King’s College London, UK)
  • Arumugam Nallanathan (Queen Mary University of London, UK)

Motivation and Context

Nowadays, the mobile network no longer just connects people but is evolving into billions of devices, such as sensors, extended reality, controllers, machines, autonomous vehicles, drones, people and things with each other and then achieves information and Intelligence. From a planning and optimization perspective on the mobile network, this means that we also need a lot more flexibility to address these future needs. Next-generation (B5G/6G) mobile networks are characterized by three key features: heterogeneity, in terms of technology and services, dynamics, in terms of rapidly varying environments and uncertainty, and size, in terms of number of users, nodes, and services. The need for smart, secure, and autonomic network design has become a central research issue in a variety of applications and scenarios. Intelligence (AI) and future wireless networks have attracted intense interest from both academia and industry to potentially improve spatial reuse and coverage, thus allowing cellular systems to achieve higher data rates, while retaining the seamless connectivity and mobility of cellular networks. However, considering the severe inter-tier interference and limited cooperative gains resulting from the constrained and non-ideal transmissions between adjacent base stations, a new paradigm for improving both spectral efficiency and energy efficiency through suppressing inter-tier interference and enhancing the cooperative processing capabilities is needed in the practical evolution of AI-based future mobile networks. This tutorial will identify and discuss technical challenges and recent results related to the AI-based future mobile networks. The tutorial will introduce future mobile networks and AI, discuss about the future mobile networks’ architecture, AI-based resource management, PHY and MAC layer design and optimization with AI and providing a future outlook of AI-based future wireless networks.

Structure and Content

Part I – Overview of Future Wireless Networks and AI

  • RAN Evolutions: Brief introduction of 6G, and its potential evolution.
  • Introduction of AI based Future Wireless Networks: Features, definitions, challenges, and state of the art.
  • System architecture: Fronthaul, Fog/cloud computing, heterogeneous networks, performance metrics

Part II – AI based Resource Management in Future Wireless Networks

  • Artificial Intelligence based resource allocation in ultra-dense networks
  • Deep neural network based power control for NOMA networks
  • Cross layer optimization in AI based future wireless networks
  • User association and power allocation using deep learning

Part III – AI based Interference Management in Future Wireless Networks

  • Learning based interference mitigation
  • AI based interference mitigation and handover management
  • Coexistence of Wi-Fi and UDN with LTE-U
  • Incomplete CSI based resource optimization in SWIPT

Part IV – AI-based MAC Control in IoT Networks

  • Traffic load prediction in framed-ALOHA networks
  • Deep reinforcement learning for NB-IoT Networks
  • A decoupled learning strategy for massive IoT Networks

Part V – AI solutions in Wireless Virtual Reality Networks

  • Learning-based Prediction, Rendering, and Association Optimization
  • Learning-based Correlation-aware Cooperative Multigroup Broadcast 360° Video Delivery Network

Part VI – Outlook of AI-based Future Wireless Networks

  • Evolution of AI-based Future Wireless Networks: Future research challenges
Go to Top