WDM-VLC Receiver Sensors: Large-Scale Filter-Array Detectors With Optimized Selection Combining Methods

Due to the continuous improvement of solid-state lighting technologies, multispectrum, multicolor LEDs have been used as transmitters for wavelength division multiplexing visible light communication (WDM-VLC) techniques. Although the multispectrum LEDs can emit light with different wavelengths for multichannel transmission, the overlapped spectra across the WDM channels could cause high interband multichannel interferences (IBMCI). Accordingly, the large-scale filter-array detectors (LFDs) have been adopted to WDM-VLC techniques and shown high capability for the IBMCI rejection using the maximum signal-to-interference-plus-noise ratio (SINR) combining (MSC) method. However, in the MSC method, it requires 100% filter array usage ratio, which would potentially result in costly computational complexity. In this paper, we propose two approaches, namely, conjugate gradient (CG) algorithm approach and binary particle swarm optimization (BPSO) algorithm approach, for sparse filter selection. According to both simulation and experimental results, the proposed CG algorithm approach can significantly reduce the filter array usage ratio by up to 70% while maintaining a good detector output SINR. The proposed BPSO algorithm approach can provide a near optimal detector output SINR while maintaining a good reduction of filter array usage ratio.