Least Minimum Symbol Error Rate Based Post-Distortion for VLC Using Random Fourier Features

Mitra, Rangeet; Jain, Sandesh; Bhatia, Vimal

doi:10.1109/lcomm.2020.2968026

Cited by 30 publications

(14 citation statements)

References 17 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The aforementioned RKHS/RFF based post-distorters (namely kernel least mean square (KLMS) [19], kernel minimum symbol error rate (KMSER) [9], KLMS with decision feedback equalizer (KLMS-DFE) [20], and RFF-KMSER-DFE [18]) are based on stochastic gradient descent (SGD) paradigms that have lower computational complexity, however higher steady-state mean square error (MSE). Further, the aforesaid SGD based post-distortion algorithms considers only instantaneous error for formulation of loss function, and hence delivers suboptimal performance under high date rate regime [21]. Furthermore, to achieve a trade-off between the convergence rate, and the MSE, the existing KLMS-DFE, and RFF-KMSER-DFE based post-distorters consider adaptive step-size, which itself are sensitive to the choice of various hyperparameters as given in [18, eq.…”

Section: A Related Workmentioning

confidence: 99%

“…To further improve the convergence performance of RKHS based algorithms, least-squares based kernel recursive least squares (KRLS) [14], least KMSER [21], and kernel recursive maximum Versoria criterion (KRMVC) [22] algorithms have been proposed in the literature for a VLC post-distortion algorithm. These algorithms consider all the previous error terms for objective function formulation, and hence outperforms the existing SGD based approaches.…”

Section: A Related Workmentioning

confidence: 99%

“…However, the performance of RKHS/RFF based postdistorters [18], [21], and the Gaussian kernel based DNN (GK-DNN) equalizers [12] are sensitive to the choice of appropriate kernel-width, and results in degraded performance upon improper choice of kernel-width. There are several methods in the signal processing literature to select proper kernelwidth like: (a) Silverman's and Scott's rule of thumb [14], (b) adaptive techniques [26], [27] for learning kernel-width based on various criteria such as MMSE, minimum symbol error rate (MSER) etc., and (c) Bayesian leaning methods [28].…”

Section: A Related Workmentioning

confidence: 99%

See 2 more Smart Citations

Kernel Recursive Maximum Versoria Criterion Based Post-Distorter for VLC Using Kernel-Width Sampling

et al. 2022

Self Cite

View full text Add to dashboard Cite

Visible light communication (VLC) has emerged as a potential candidate for next generation wireless communication systems. However, nonlinear characteristics of light emitting diode (LED), user-mobility, and DC-bias fluctuations are the major factors that limit the throughput of a VLC link, and makes the overall additive distortion as non-Gaussian distributed. To mitigate this non-Gaussian noise processes encountered in VLC systems due to LED nonlinearity, and user-mobility, recently a random Fourier features (RFF) based kernel recursive maximum Versoria criterion (KRMVC) based post-distortion algorithm is proposed, which delivers better performance as compared to the classical polynomial series, and kernel recursive least squares (KRLS) algorithms due to the incorporation of higher order statistics of error. However, the performance of RFF-KRMVC algorithm is sensitive to the choice of kernel-width, and results in approximation errors due to imperfect choice of kernelwidth. This paper proposes a novel RFF-KRMVC algorithm using a kernel-width sampling (KWS) technique called as RFF-KWS-KRMVC, which implements the post-distortion under a hyperparameter-free finite memory budget. Furthermore, analytical expressions for mean square error, and error rate are quantified for the proposed RFF-KWS-KRMVC post-distorter, and corroborated by Monte-Carlo simulations performed over standard VLC channel models.

show abstract

Section: A Related Workmentioning

confidence: 99%

Section: A Related Workmentioning

confidence: 99%

Section: A Related Workmentioning

confidence: 99%

See 1 more Smart Citation

Kernel Recursive Maximum Versoria Criterion Based Post-Distorter for VLC Using Kernel-Width Sampling

et al. 2022

Self Cite

View full text Add to dashboard Cite

show abstract

“…Moreover, QKD is an emerging quantum technology for securing existing and future communication networks. Hence, QKD can be used for providing security in free space optics (FSO) communication systems [229], visible light communications [230], [231], [232], [233], [234], [235], [236], Unmanned Aerial Vehicles (UAVs) communications [237], [238], and THz communications [239], [240].…”

Section: G Other Applications and Research Directionsmentioning

confidence: 99%

Quantum Key Distribution Secured Optical Networks: A Survey

Sharma

Agrawal

Bhatia

et al. 2021

IEEE Open J. Commun. Soc.

Self Cite

View full text Add to dashboard Cite

Increasing incidents of cyber attacks and evolution of quantum computing poses challenges to secure existing information and communication technologies infrastructure. In recent years, quantum key distribution (QKD) is being extensively researched, and is widely accepted as a promising technology to realize secure networks. Optical fiber networks carry a huge amount of information, and are widely deployed around the world in the backbone terrestrial, submarine, metro, and access networks. Thus, instead of using separate dark fibers for quantum communication, integration of QKD with the existing classical optical networks has been proposed as a cost-efficient solution, however, this integration introduces new research challenges. In this paper, we do a comprehensive survey of the state-of-the-art QKD secured optical networks, which is going to shape communication networks in the coming decades. We elucidate the methods and protocols used in QKD secured optical networks, and describe the process of key establishment. Various methods proposed in the literature to address the networking challenges in QKD secured optical networks, specifically, routing, wavelength and time-slot allocation (RWTA), resiliency, trusted repeater node (TRN) placement, QKD for multicast service, and quantum key recycling are described and compared in detail. This survey begins with the introduction to QKD and its advantages over conventional encryption methods. Thereafter, an overview of QKD is given including quantum bits, basic QKD system, QKD schemes and protocol families along with the detailed description of QKD process based on the Bennett and Brassard-84 (BB84) protocol as it is the most widely used QKD protocol in the literature. QKD system are also prone to some specific types of attacks, hence, we describe the types of quantum hacking attacks on the QKD system along with the methods used to prevent them. Subsequently, the process of point-to-point mechanism of QKD over an optical fiber link is described in detail using the BB84 protocol. Different architectures of QKD secured optical networks are described next. Finally, major findings from this comprehensive survey are summarized with highlighting open issues and challenges in QKD secured optical networks.

show abstract

“…It is noted that while there are several extensions from x possible, the RFF based extension outperforms classical polynomial kernel based extensions[33] …”

mentioning

confidence: 99%

Random Fourier Feature Based Deep Learning for Wireless Communications

Mitra¹,

Kaddoum²

2021

Preprint

Self Cite

View full text Add to dashboard Cite

Deep-learning (DL) has emerged as a powerful machine-learning technique for several classic problems encountered in generic wireless communications. Specifically, random Fourier Features (RFF) based deep-learning has emerged as an attractive solution for several machinelearning problems; yet there is a lacuna of rigorous results to justify the viability of RFF based DL-algorithms in general. To address this gap, we attempt to analytically quantify the viability of RFF based DL. Precisely, in this paper, analytical proofs are presented demonstrating that RFF based DL architectures have lower approximationerror and probability of misclassification as compared to classical DL architectures. In addition, a new distributiondependent RFF is proposed to facilitate DL architectures with low training-complexity. Through computer simulations, the practical application of the presented analytical results and the proposed distribution-dependent RFF, are depicted for various machine-learning problems encountered in next-generation communication systems such as: a) line of sight (LOS)/non-line of sight (NLOS) classification, and b) message-passing based detection of low-density parity check codes (LDPC) codes over nonlinear visible light communication (VLC) channels. Especially in the low training-data regime, the presented simulations show that significant performance gains are achieved when utilizing RFF maps of observations. Lastly, in all the presented simulations, it is observed that the proposed distributiondependent RFFs significantly outperform RFFs, which make them useful for potential machine-learning/DL based applications in the context of next-generation communication systems.

show abstract

Least Minimum Symbol Error Rate Based Post-Distortion for VLC Using Random Fourier Features

Cited by 30 publications

References 17 publications

Kernel Recursive Maximum Versoria Criterion Based Post-Distorter for VLC Using Kernel-Width Sampling

Kernel Recursive Maximum Versoria Criterion Based Post-Distorter for VLC Using Kernel-Width Sampling

Quantum Key Distribution Secured Optical Networks: A Survey

Random Fourier Feature Based Deep Learning for Wireless Communications

Contact Info

Product

Resources

About