Mitra, R.Miramirkhani, F.Bhatia, V.Uysal, Murat2021-02-232021-02-232020-100018-9545http://hdl.handle.net/10679/7345https://doi.org/10.1109/TVT.2020.3018740Vehicular visible light communications (VLC) has emerged as a viable supplement for high speed next-generation vehicle to vehicle (V2V) communication systems. However, performance of a V2V-VLC link is impaired due to nonlinear transfer-characteristics of light emitting diodes (LEDs), and inter-symbol interference (ISI). In this article, a low-complexity least-squares based post-distortion algorithm is formulated over reproducing kernel Hilbert space (RKHS) for a multi-hop V2V-VLC link. The impairments encountered in V2V-VLC channels are mitigated in RKHS by a minimum symbol error-rate post-distorter using a low dimensional approximation of random Fourier features (RFF) (which is a soft approximation of the feature-map to RKHS), that facilitates computationally simple post-distortion under finite memory-budget. The convergence and the BER-performance of the proposed post-distorter is analyzed over realistic V2V VLC channels obtained via ray-tracing. From the analysis, and the presented computer-simulations, the proposed post-distorter is found to exhibit equivalent convergence characteristics and error-rate over reasonable distances, with much lower computational complexity.engrestrictedAccessarticle6910118001181000058376760009110.1109/TVT.2020.3018740RFFAutoencodersV2VVLCRKHS2-s2.0-85095721812