Optimal Resource Management for NOMA-Based Visible Light Communication Systems With Shot Noise

In this paper, the problem of power allocation is investigated for downlink visible light communication (VLC) systems with non-orthogonal multiple access. In the system, both input-dependent shot noise and input-independent Gaussian noise are considered due to the properties of practical VLC channels. We consider two scenarios for the transmitter with a single light emitting diode (LED) and multiple LEDs, respectively. For single-LED scenarios, the problem is posed as a joint optimization problem of alternating current and direct current power control, whose goal is to maximize the minimum signal-to-interference-plus-noise-ratio (SINR) of all users. Although the original problem is non-convex, a geometric programming based algorithm is proposed to acquire the global optimal solution. For multi-LED scenarios, the transmit and dimming beamforming vectors are jointly optimized to maximize the minimum SINR of all users. A joint optimization algorithm that alternately optimizes transmit and dimming beamforming vectors is proposed. Particularly, the transmit beamforming problem is converted into a second-order cone programming problem through the successive convex approximation method. The dimming beamforming problem is transformed into a feasibility problem utilizing a low-complexity bisection algorithm. Numerical results reveal the superior performance of the proposed algorithm compared with its orthogonal multiple access counterpart and conventional schemes without shot noise.