Nonlinear Distortion-Aware Channel Estimation Strategies for IRS-Aided MIMO Systems

The article addresses the challenges posed by nonlinear distortions in Intelligent Reflecting Surface (IRS)-assisted Multiple-Input Multiple-Output (MIMO) systems, which are often overlooked in conventional channel estimation strategies. A distortion-aware channel estimation framework is proposed for IRS-assisted MIMO systems operating under spatially correlated channel perturbations. The framework employs Minimum Mean Square Error (MMSE) and Least Squares (LS) estimators to evaluate the impact of nonlinear distortions arising from non-Independent and Identically Distributed (non-IID) Gaussian noise and non-uniform phase variations. Monte Carlo simulations are conducted to analyses Bit Error Rate (BER) performance, considering the influence of IID/non-IID Gaussian configurations on the LS and MMSE estimation methods in realistic fading environments. Integration of IRS elements yields Signal-to-Noise Ratio (SNR) enhancements by ≈ 5 dB in Rician fading and ≈ 3 dB under correlated Rician fading conditions, accounting for both IID and non-IID Gaussian variations. Furthermore, an optimized MMSE variant is introduced, incorporating statistically aligned training sequence and power allocation strategies, which significantly enhances Normalized Mean Square Error (NMSE) performance in nonlinear distorted channels. Comparative benchmarking confirms the scalability, and practical relevance of the proposed framework, highlighting its suitability for next-generation wireless communication systems.