Deep learning-based channel estimation for MIMO system | |
| Author | Pornpetch Gansaree |
| Call Number | AIT RSPR no.TC-25-02 |
| Subject(s) | MIMO systems Deep learning (Machine learning) |
| Note | A research study submitted in partial fulfillment of the requirements the degree of Master of Engineering in Telecommunications |
| Publisher | Asian Institute of Technology |
| Abstract | In recent times, significant advancements have occurred within wireless communication technologies. Fifth-generation (5G) wireless systems have gained broad adoption to address the increasing need for high-speed data transmission. To boost network capabilities, sophisticated methodologies like Multiple-Input Multiple Output (MIMO) and Orthogonal Frequency Division Multiplexing (OFDM) are implemented. Both technologies require efficient and reliable channel estimation algorithms to improve network capacity, spectral efficiency, and overall performance. Accurate channel estimation is essential as it directly influences communication network performance and efficiency. However, conventional channel estimation approaches struggle to operate effectively in dynamic and unpredictable nature of modern wireless environments. Recently, developments in Artificial Intelligence (AI), particularly through deep learning approaches, have demonstrated potential in addressing these complicated challenges.This research primarily aims to create an AI-enhanced channel estimation framework that improves the accuracy, efficiency, and adaptability of MIMO-OFDM systems in 5G. By leveraging deep learning methodologies, this study addresses the limitations inherent in conventional techniques, such as Least Squares (LS) and Linear Minimum Mean Square Error (LMMSE), which often face difficulties related to computational complexity, noise sensitivity, and dependence on prior statistical knowledge.This work explores and compares the efficiency of two deep learning models: a Denoising Autoencoder (DAE) and Convolutional Neural Network (CNN), each designed to estimate MIMO channels under realistic wireless conditions, including Additive White Gaussian Noise (AWGN), Rayleigh, and Rician fading. Utilizing the DeepMIMO dataset, the models are trained and evaluated, ensuring that they reflect practical deployment scenarios. By effectively integrating the strengths of both models, this research highlights the ability of deep learning to deliver accurate and efficient channel estimation in dynamic 5G environments, while offering a scalable foundation for future wireless systems such as massive MIMO. |
| Year | 2025 |
| Type | Research Study Project Report (RSPR) |
| School | School of Engineering and Technology |
| Department | Department of Information and Communications Technologies (DICT) |
| Academic Program/FoS | Telecommunications (TC) |
| Chairperson(s) | Attaphongse Taparugssanagorn |
| Examination Committee(s) | Chaklam Silpasuwanchai;Chantri Polprasert |
| Scholarship Donor(s) | Royal Thai Government;His Majesty the King’s Scholarships |
| Degree | Research Studies Project Report (M. Eng.) - Asian Institute of Technology, 2025 |