Dynamic channel estimation in large-scale massive MIMO systems with intelligent reflecting surfaces: Leveraging Khatri-Rao factorization and bilinear alternating least squares

In large-scale massive MIMO systems with intelligent reflecting surfaces (IRS), dynamic channel estimation (CE) is essential for optimizing the system performance and ensuring reliable communication. Traditional channel estimation techniques are not suitable for IRS-assisted systems due to the unique characteristics of Intelligent Reflecting Surfaces channels. To address the channel estimation problem in such dynamic environments, this paper introduces two novel channel estimation methods: Khatri-Rao Factorization (KRF) and Bilinear Alternating Least Squares (BALS). The first method uses KRF to efficiently solve rank-1 matrix approximation problems with a closed-form solution. The second method employs an iterative alternating estimation scheme. By disentangling these key channel matrices' estimates, both methods provide more accurate and robust channel estimation, essential for optimizing communication system performance in challenging environments. The proposed CE-KRFBALS-MIMO method is evaluated under performance metrics like Bit error rate (BER), Signal Noise Ratio (SNR), Normalized Mean Square Error (NMSE) Spectral Efficiency (SE), and Computational Complexity.