Intelligent Wireless Channel Prediction: A Sensitivity Analysis for Future Networks
Problem Statement
AI has the potential to revolutionize wireless communication by enabling intelligent scheduling of resources in base stations of cellular networks. Effective scheduling depends on the ability to manage traffic variations and adapt to wireless channel fluctuations. This thesis explores the exciting possibility of using AI to predict wireless channel quality before transmission. Accurate prediction could lead to better utilization of communication resources and enhanced network performance.
To make meaningful progress, this research will analyze the sensitivity of prediction error, mobility, scalability, and other key parameters on the usefulness of wireless channel prediction across various environments such as highways (intelligent transportation), urban areas (smart cities), rural areas (smart agriculture) and indoor (smart homes/industries/offices) settings. Sensitivity analysis will help uncover the limits and potential of AI-driven predictions in different conditions.
Building on the work of Sabari Nathan Anbalagan, who performed a similar study for industrial indoor environments, this research offers students the opportunity to explore cutting-edge applications of AI in a domain with immense technological relevance.
Tasks
· Use MATLAB to simulate wireless communication in highway, urban, and indoor environments.
· Analyze the impact of prediction error, mobility, scalability, and other relevant parameters on the effectiveness of AI-based channel quality prediction.
· Compare and contrast the sensitivity across different environments.
· Use findings to provide actionable insights into the feasibility of AI-driven scheduling for wireless communication systems.
Work Distribution
· 30% Theory: Understand wireless communication principles, AI techniques for prediction, and the role of environment-specific factors.
· 50% Simulations: Use QuaDRiGa to model environments and test AI algorithms under varying conditions.
· 20% Writing: Document findings, focusing on the significance of results and potential real-world applications.
Contact:
Sabari Nathan Anbalagan (s.n.anbalagan@utwente.nl)