Channel modelling to ensure high link reliability

• Main Author: Lam, Teresa Xin Yuan
• Other Authors: Lee Yee Hui
• Format: Final Year Project
• Language: English
• Published: Nanyang Technological University 2024
• Subjects: Engineering; Channel modelling
• Online Access: https://hdl.handle.net/10356/176399
• Institution: Nanyang Technological University

This study explores the application of RadioUNet, a Convolutional Neural Network (CNN) architecture, for high-reliability channel estimation. With an emphasis on high-reliability wireless communication, this research examines the usage of RadioUNet, a CNN architecture, for channel estimation. RadioUNet was selected for its creative approach to radio map generation and signal strength estimates, utilising UNet's upsampling and downsampling for picture segmentation. The main objective of the project is to gather Received Signal Strength (RSS) data from several transmitters and receivers located across NTU, train RadioUNet with simulated scenarios, and then use that data to create precise signal strength images on NTU's map. The objective is to increase the model's accuracy under practical circumstances, hence enhancing the dependability of wireless channel estimation. The output of the model will be validated against real data to create a framework for effective and trustworthy channel modelling in comparable circumstances.