Indoor visible light communication (VLC) using light-emitting diodes (LEDs) faces critical performance challenges due to LED nonlinearity, limited modulation bandwidth, and multipath-induced inter-symbol interference. This study evaluates carrierless amplitude and phase (CAP) modulation across single-input single-output (SISO), spatial multiplexing (SMP), and spatial modulation (SM) configurations, employing linear least mean square (LMS) and nonlinear Legendre functional link artificial neural network (Le-FLANN) and Volterra equalization techniques. Under realistic LED operating conditions and spatial misalignment, Le-FLANN achieves significant signal-to-noise-ratio per bit ( E b / N 0 ) improvements at bit error rate (BER) = 10 −4 : 0.9 dB in SISO, 2.9 dB in SMP, and 1.5 dB in SM, compared to LMS. SMP delivers an aggregate throughput of 24 Mbps by transmitting two independent 12 Mbps CAP streams in parallel over a 4.5 MHz LED bandwidth. These findings demonstrate the practical value of structured nonlinear equalization in bandwidth-constrained VLC links, offering enhanced resilience to LED-induced distortion and spatial channel impairments.
Ashouri et al. (Fri,) studied this question.