MMV-Aided Bayesian Learning for Simultaneous Sparse Channel Estimation in Visible Light DCO-OFDM Systems

A multipath channel impulse response (CIR) estimator is introduced, utilizing the simultaneous sparsity across multiple measurement vectors (MMV) for direct current-biased optical OFDM (DCO-OFDM) visible light communication (VLC) systems. The multipath CIR model incorporates both diffusive and specular reflections in the VLC channel. The system model for the DCO-OFDM is formulated, followed by traditional and compressive sensing (CS)-based channel estimation (CE) methods, such as simultaneous sparse orthogonal matching pursuit (SOMP) and Bayesian learning (BL). An advanced estimation method using MMV-based Bayesian learning (MBL) is proposed, reducing pilot overhead by leveraging simultaneous sparsity. Benchmarks like the Oracle-minimum mean squared error (O MMSE) estimator and the Bayesian Cramer Rao lower bound (BCRLB) are derived for performance comparison. Simulation results show that the proposed MMV-based MBL method outperforms other techniques in normalized mean-square-error (NMSE), pilot overhead, and bit error-rate (BER), closely match the benchmarks derived.