Research on channel estimation based on joint perception and deep enhancement learning in complex communication...

Area of Science:

• Wireless Communication Engineering
• Artificial Intelligence in Communications
• Signal Processing

Background:

• Wireless systems face challenges like multipath propagation and interference, degrading performance.
• Accurate channel estimation is crucial for maintaining communication quality and data rates.
• Intelligent Reflective Surfaces (IRS) offer potential for optimizing wireless channels.

Purpose of the Study:

• To develop an intelligent channel estimation model for IRS-assisted wireless systems.
• To enhance channel estimation accuracy and robustness in dynamic environments.
• To leverage deep learning and reinforcement learning for adaptive channel optimization.

Main Methods:

• Proposed an intelligent channel estimation model fusing CNN and GRU features.
• Utilized Deep Deterministic Policy Gradient (DDPG) reinforcement learning for adaptive optimization.
• Developed a Channel Reconstruction Prediction and Generation Network (CRPG-Net) for feature extraction.

Main Results:

• The proposed model achieved significantly lower channel estimation error compared to LS and LMMSE methods.
• Demonstrated superior accuracy and robustness across public datasets and real-world scenarios.
• The reinforcement learning component enabled continuous optimization in dynamic channel conditions.

Conclusions:

• The CNN-GRU-DDPG based CRPG-Net offers an innovative and effective approach to IRS-assisted channel estimation.
• This method significantly enhances communication quality and system performance in complex wireless environments.
• The intelligent model provides a robust solution for adapting to changing channel states.