OFDM Bitloading in Distributed MIMO OWC Using Power-Constrained LEDs

Abstract

The combined use of distributed multiple-input multiple-output (D-MIMO) and orthogonal frequency division multiplexing (OFDM) with adaptive bit and power loading is seen as a powerful tool to overcome limitations of the optical wireless communication (OWC) channel and of light emitting diodes (LED)s. Although extensive research into efficient MIMO algorithms has been performed for communications via radio, these may not be indiscriminately re-used in LED-based OWC systems. The main reasons are that the OWC channel has different properties, the LEDs have power-constraints, and eyesafety regulations apply. For adoption in a practical system, the computational complexity of the algorithms must also be taken into account. Fortunately, physical properties of the LED D-MIMO channel allow us to reduce signalling overhead, estimation effort and complexity. This paper presents, compares, and discusses both the performance and the computational cost of several new power-loading strategies to enhance the throughput of LED-based OFDM D-MIMO OWC systems respecting the individual power constraints of the light sources. To cope with the low-pass nature of LEDs, we optimize how much power to load on each MIMO spatial stream, on each sub-carrier and what bandwidths to use. The CVX toolbox for convex optimization gives us an optimum but it is orders of magnitude more complex than highly effective heuristic approaches. A simple approach stays 1 dB short of the optimum: it equally splits power among MIMO streams within the LED power constraints, and it uniformly spreads the power over a bandwidth that is optimized per stream. A similar strategy, but selecting optimized powers per stream brings us as near as 0.1 dB to the optimum. To evaluate the complexity of bit and power allocations, we investigate, optimize, and accelerate the speed of iterative schemes. The proposed approaches give new choices for chip set manufacturers and LED-based OFDM D-MIMO OWC system designers to improve performance and to reduce overhead.