Wireless Communications

Advanced research, algorithms, and 5G/6G wireless communications

Wireless communication systems are undergoing a paradigm shift towards 5G-Advanced and 6G networks, driven by Extremely Large-Scale Antenna Arrays (ELAA), near-field propagation, high-frequency spectrums, and artificial intelligence. My research focuses on developing innovative channel estimation algorithms, informed system identification frameworks, and deep learning models for next-generation physical layers.

Our solutions span theoretical performance limits (Cramér-Rao Bounds), robust channel identification (InSI framework, semi-blind equalizers), and deep learning-based channel estimators (RACNN, ISDNN).

Our Research Approaches
01 System Identification

Informed System Identification (InSI) Toolbox

Semi-blind Equalization · Induced Properties · Prior Knowledge

Conventional blind channel estimation suffers from high sample requirements and phase ambiguities. InSI is our developed MATLAB framework that incorporates high-level prior information, induced signal properties, and semi-blind processing to overcome these limitations.

InSI evaluates Cramér-Rao Bounds (CRB) across FIR, specular, and structured channel models, providing optimal estimation accuracy for complex MIMO-OFDM communication systems.

Son, MSc Thesis (2023); Son et al., APSIPA ASC (2023); Repository: InSI.
Informed System Identification Framework
Informed System Identification framework & Cramér-Rao bound analysis.
02 6G Near-Field

Near-Field Communication & Ray-Tracing Channel Models

Extremely Large Antenna Arrays (ELAA) · RACNN Deep Learning

In 6G networks, Extremely Large-Scale Antenna Arrays (ELAA) expand the near-field region where spherical wavefronts dominate. We develop ray-tracing propagation models and deep learning architectures to tackle near-field channel estimation.

Our proposed Residual Attention Convolutional Neural Network (RACNN) captures spatial spherical correlations, achieving superior channel estimation accuracy with significantly reduced training complexity.

Khang, Son et al., VNU JCSCE (2026); Lam, Son et al., CITA (2025); Repository: RACNN.
6G Near-Field Ray Tracing & RACNN
Spherical wavefront ray-tracing & RACNN deep learning for 6G near-field.
03 Massive MIMO

Deep Neural Networks & 3D Antenna Array Geometry

ISDNN Deep Learning · Array Geometry (ULA, UCyA) Analysis

Massive MIMO systems require accurate estimation of high-dimensional channel matrices under varying SNR levels. We design ISDNN, a customized deep neural network architecture for fast and robust Massive MIMO channel estimation.

Additionally, we perform comprehensive Cramér-Rao Bound (CRB) impact analyses on 3D antenna array geometries (Uniform Linear, Uniform Cylindrical Arrays), establishing physical array design guidelines.

Son, Lam et al., HaUI JST (2024); Son et al., IEEE SSP (2023).
Massive MIMO DNN & Antenna Geometry
ISDNN neural network architecture & 3D antenna array geometry performance.
Publications

Journals

  1. Nguyen The Khang, Do Hai Son, Tran Trong Duy, Le Thanh Trung, Tran Thi Thuy Quynh, Karim Abed-Meraim, Merouane Debbah, and Nguyen Linh Trung. Ray-Tracing Channel Models for Near-Field Communication. VNU Journal of Science: Computer Science and Communication Engineering, 42(3):1-24, June 2026.
  2. Do Hai Son, Vu Tung Lam, and Tran Thi Thuy Quynh. ISDNN: A Deep Neural Network for Channel Estimation in Massive MIMO systems. Hanoi University of Industry Journal of Science and Technology, 60(11):48-54, November 2024.

Conferences & Workshops

  1. Vu Tung Lam, Do Hai Son, Tran Thi Thuy Quynh, and Le Trung Thanh. RACNN: Residual Attention Convolutional Neural Network for Near-Field Channel Estimation in 6G Wireless Communications. In Conference on Information Technology and its Applications (CITA), Phnom Penh, Cambodia, July 2025.
  2. Hai Pham Anh, Tran Trong Duy, Do Hai Son, Karim Abed-Meraim, and Nguyen Linh Trung. FlowEKF: Flow-based Extended Kalman filter. In Asia Pacific Signal and Information Processing Association Annual Summit and Conference (APSIPA ASC), Singapore, October 2025.
  3. Do Hai Son, Karim Abed-Meraim, Tran Trong Duy, Nguyen Linh Trung, and Tran Thi Thuy Quynh. On the Semi-Blind Mutually Referenced Equalizers for MIMO Systems. In Asia Pacific Signal and Information Processing Association Annual Summit and Conference (APSIPA ASC), Taipei, Taiwan, November 2023.
  4. Do Hai Son and Tran Thi Thuy Quynh. Impact Analysis of Antenna Array Geometry on Performance of Semi-blind Structured Channel Estimation for massive MIMO-OFDM systems. In IEEE Statistical Signal Processing Workshop (SSP), Hanoi, Vietnam, July 2023.

Books & Theses

  1. Do Hai Son. Nghiên cứu nhận dạng hệ thống với tri thức mới cho hệ thống truyền thông MIMO kích thước lớn. Luận văn Thạc sĩ, Trường Đại học Công nghệ, Đại học Quốc gia Hà Nội, Tháng 6/2023.