Distributed Power Allocation for Multiuser MISO Indoor Visible Light Communications

In multiuser MISO (multiple-input single output) visible light communication (VLC) systems, it is essential to find effective means for allocating the transmitted power from the various LEDs fairly among users that are randomly dispersed in large indoor environments. Distributed algorithms are a natural choice because, unlike centralized power allocation algorithms, distributed schemes do not require extensive computation complexity and computation time. In this paper we propose and compare two distributed algorithms: a partial jointly optimized power allocation and a weighted distributed power allocation. The system uses code division multiple access (CDMA) plus optical on- off keying (OOK) with a minimum mean squared error (MMSE) receiver to diminish the effect of multiple-access interference in the indoor VLC system. We compare the performance of the proposed distributed algorithms with an optimal centralized approach. From the numerical results, both distributed algorithms proposed in this paper have less than a 2 dB power penalty compared with the centralized approach, and the computational complexity can be reduced by 75% for a small indoor environment and by 93% for a large indoor environment.