Adaptive Channel Estimation for Semi-Passive IRS with Optimized Sensor Deployment.

To achieve optimal passive beamforming gains from Intelligent Reflective Surfaces (IRS), accurate Channel State Information (CSI) acquisition is required. However, the IRS, with numerous passive devices, lacks the ability to process signals, resulting in considerable challenges in obtaining accurate CSI. Based on the semi-passive IRS, this paper proposes a compressed sensing channel estimation algorithm without knowing the path number of channel, which improves the accuracy of channel estimation. Furthermore, a particle swarm optimization (PSO)-based deployment scheme for active sensors in the semi-passive IRS is developed. Numerical simulations confirm the effectiveness, demonstrating a reduction in Normalized Mean Square Error (NMSE) and improved channel estimation with fewer pilot symbols, thereby minimizing estimation overhead.