Capacity-Maximized Transmitter Precoding for MU MIMO VLC Systems With Bounded Channel Uncertainties

Considering imperfect channel state information (CSI), we study robust precoding design for multiuser multiple-input-multiple-output (MIMO) indoor visible-light communication systems, where one transmitter with several light-emitting diodes sends messages to multiple receivers each with several photodiodes. We model the imperfect CSI with bounded channel uncertainties by worst-case model, aiming to maximize the worst-case sum capacity of all receivers via optimizing the precoders at the transmitter under optical power constraints. The robust precoding problem is nonconvex and thus hard to solve. To overcome this difficulty, we first separate the logarithm function and the precoders in the objective function, then employ the sign-definiteness lemma to eliminate the channel uncertainties, and finally develop an effective iterative optimization algorithm to compute the precoders. Simulation results are provided to demonstrate the effectiveness of the proposed scheme.