Channel Estimation Algorithm with Fuzzy-Enhanced Compressed Sensing for Indoor MIMO VLC

The channel estimation approaches for indoor multiple-input multiple-output (MIMO) visible light communication (VLC) systems based on compressed sensing (CS) theory, can effectively reduce the pilot overhead for channel training and the computational cost. However, conventional CS reconstruction techniques, such as the sparse adaptive matching pursuit (SAMP), are unable to achieve a satisfactory balance between accuracy and efficiency. To address this issue, we propose an algorithm that combines the fuzzy control strategy with SAMP, namely FC-SAMP. This algorithm utilizes fuzzy rules to simulate human decision-making processes. It dynamically adapts the step size by considering the iterative residuals and their variation rate, to achieve an efficient and accurate estimate for sparse channel state information. The simulation results show that the proposed FC-SAMP outperforms the orthogonal matching pursuit (OMP), the SAMP, and other variable step size algorithms in indoor MIMO VLC systems, in terms of the convergence rate and estimation error.