PaperlayerView exact revision on arXiv

Source-linked AI summary

Physical Layer Security for Visible Light Communication Systems: A Survey

Mohamed Amine Arfaoui, Mohammad Dehghani Soltani, Iman Tavakkolnia, Ali Ghrayeb, Chadi Assi, Majid Safari, Harald Haas

arXiv:1905.11450v1eess.SP

TL;DR

VLC security remains an open concern because its channels are open and broadcasting, while a unified PLS survey spanning information-theoretic and signal-processing perspectives is missing. This paper reviews PLS-VLC studies across channel, signaling, architecture, CSI, deployment, and secrecy-performance dimensions, and identifies research directions including realistic channel models, mobility, orientation, and VLC/RF integration.

  • Problem

    A comparative and unified survey of physical layer security for VLC, covering information-theoretic and signal-processing perspectives, is missing despite VLC’s open broadcasting channels and relevance to 5G.

  • Method

    The paper synthesizes PLS-VLC research by organizing studies according to channel characteristics, input distributions, system architectures, users and eavesdroppers, CSI, deployment geometry, and signaling schemes.

  • Results

    The survey covers PLS-VLC channel models, signaling and precoding strategies, network configurations, secrecy capacity, and information rates, while identifying unresolved indoor-security problems and realistic-channel research directions.

  • Takeaways & Limitations

    Future PLS-VLC research should consider discrete signaling, realistic mobility- and orientation-aware channel models, generalized multiuser MIMO settings, and hybrid VLC/RF systems.

  • Takeaways & Limitations

    Several important settings remain insufficiently studied, including discrete-input MU-MISO systems and MU-MIMO VLC broadcast channels with colluding eavesdroppers and imperfect CSI.

Abstract

from arXiv · show

Due to the dramatic increase in high data rate services and in order to meet the demands of the fifth-generation (5G) networks, researchers from both academia and industry are exploring advanced transmission techniques, new network architectures and new frequency spectrum such as the visible light spectra. Visible light communication (VLC) particularly is an emerging technology that has been introduced as a promising solution for 5G and beyond. Although VLC systems are more immune against interference and less susceptible to security vulnerabilities since light does not penetrate through walls, security issues arise naturally in VLC channels due to their open and broadcasting nature, compared to fiber-optic systems. In addition, since VLC is considered to be an enabling technology for 5G, and security is one of the 5G fundamental requirements, security issues should be carefully addressed and resolved in the VLC context. On the other hand, due to the success of physical layer security (PLS) in improving the security of radio-frequency (RF) wireless networks, extending such PLS techniques to VLC systems has been of great interest. Only two survey papers on security in VLC have been published in the literature. However, a comparative and unified survey on PLS for VLC from information theoretic and signal processing point of views is still missing. This paper covers almost all aspects of PLS for VLC, including different channel models, input distributions, network configurations, precoding/signaling strategies, and secrecy capacity and information rates. Furthermore, we propose a number of timely and open research directions for PLS-VLC systems, including the application of measurement-based indoor and outdoor channel models, incorporating user mobility and device orientation into the channel model, and combining VLC and RF systems to realize the potential of such technologies.

I. INTRODUCTION

VLC is presented as a complementary 5G technology offering high data rates and interference robustness, while its open broadcast channel still requires physical-layer security. The paper provides a unified survey of PLS-VLC research and identifies unresolved design challenges and research directions.

  • Motivation: 5G and IoT requirements include high data rates, connection density, ultra-reliable low-latency communication, and security.
  • Motivation: VLC uses visible light simultaneously for illumination and data communication, with high data rates, a large spectrum, and robustness against interference.LEDs also offer high conversion efficiency, long lifespan, low cost, and high operational speed.
  • Motivation: PLS is relevant to VLC because its real, positive optical signals and LED peak-amplitude constraints differ from RF systems.These constraints exclude unbounded Gaussian inputs from admissible capacity-achieving signaling distributions.
  • Related Work: Existing VLC security surveys did not provide a comprehensive comparative PLS treatment covering information-theoretic and signal-processing perspectives.Earlier coverage was limited in scope or depth and omitted several system-design issues, including signaling, network geometry, and CSI.
  • Contributions and Outline: This paper unifies PLS-VLC studies across channel characteristics, input distributions, architectures, users, eavesdroppers, CSI, deployment geometry, and signaling schemes.It covers SISO, MISO, MIMO, and hybrid RF/VLC systems, considering both single and multiple active users.
  • Contributions and Outline: The paper reviews VLC wiretap models and PLS methods for MIMO, MISO, SISO, and hybrid RF/VLC systems, then presents open research directions.The open-problem discussion targets more practical VLC security systems and associated secrecy-performance improvements.

D. Notations and Abbreviations

The generalized VLC wiretap model describes multi-AP transmission to multiple legitimate users and eavesdroppers, with channel, receiver, orientation, and signal constraints shaping secrecy analysis.

  • System model: The MIMO VLC system uses M APs to transmit confidential messages to multiple AUs while K EDs may intercept them.Each AU and ED can have multiple photodiodes, and the transmitter encodes messages into an M × 1 zero-mean signal vector.
  • System model: Each legitimate receiver and eavesdropper is modeled by a channel matrix and additive white Gaussian noise, with AU channel state information available through limited feedback.ED channel state information is generally unavailable because eavesdroppers may be passive or unregistered.
  • System model: The generalized model includes MIMO, MISO, and SISO VLC wiretap systems by varying the numbers of APs and photodiodes.Single-photodiode receivers yield MISO, while M = 1 yields SISO; secrecy depends on APs, users, eavesdroppers, photodiodes, and CSI availability.
  • Channel model: Most secure VLC studies consider only the line-of-sight channel component because reflected optical power is typically negligible and no readily available closed-form NLoS expression exists.This modeling choice limits how fully reflected propagation is represented in existing analyses.
  • Channel model: VLC channel descriptions use Lambertian emission, geometric distance, radiation and incidence angles, photodiode area, concentrator refractive index, and receiver field of view.The Euclidean LED–photodiode distance combines horizontal and vertical separations, while the field of view limits the incidence angle.
  • Signal constraints: LED nonlinear distortion and clipping impose a peak-amplitude constraint, making Gaussian inputs inadmissible for the corresponding information-theoretic capacity problem.The signal is bounded by the maximum allowed amplitude A at each LED fixture, although an electrical AWGN model with average power can provide a Shannon upper bound.

III. THE MIMO VLC WIRETAP SYSTEM

The MIMO VLC wiretap literature develops secrecy rates and capacity bounds for continuous and discrete signaling under amplitude constraints, then evaluates how system diversity and signaling support affect secrecy performance.

  • Single ED: For one AU and one ED, MIMO VLC secrecy capacity is formulated through the positive difference between legitimate-user and eavesdropper mutual information.The exact capacity-achieving input distribution is generally not available in closed form, motivating upper and lower bounds.
  • Single ED: Continuous log-concave inputs and a convex-concave procedure have been used to obtain achievable secrecy rates and jointly optimize the covariance matrix and signaling scheme.This approach addresses the single-AU, single-ED MIMO case.
  • Single ED: Finite-support discrete signaling provides achievable secrecy rates for practical constellations such as pulse amplitude modulation, using mass points and their probabilities.The construction relies on a relation between continuous and discrete input distributions sharing the same covariance matrix and applies to MIMO, MISO, and SISO configurations.
  • Single ED: 105 independent Monte-Carlo trials evaluate average achievable secrecy rate versus A^2 for (M, N, K) = (4, 4, 4), (4, 1, 1), and (1, 1, 1), with Q_u = 2 and Q_u = 4.The experiments fix L = 1, optimize W by brute-force search, and use noise variance E = −68.93 [dBm].
  • Single ED: Increasing transmitter light sources or receiver photodiodes increases achievable secrecy rate, while the best number of mass points depends on the amplitude-constraint operating range.The results interpret greater source or receiver multiplicity as improved system diversity and secrecy performance.
  • Single ED: The amplitude constraint ||s||∞≤A can be relaxed to the covariance trace constraint Trace(K_s)≤MA^2 to derive an upper bound on secrecy capacity.The original input support is S = [−A, A]^M, and the optimal distribution remains generally uncharacterized.

1 G1Ks

For the optimization problem in (8), the optimal covariance matrix is characterized by the active eigenvectors of H_E^T H_E. A closed-form upper bound is available for single-PD receivers, while general cases require iterative solutions.

  • The optimal covariance matrix uses the eigenvectors associated with positive eigenvalues of H_E^T H_E.
  • Closed-form upper bounds were derived for SISO and MISO VLC wiretap systems with one receiver photodetector.
  • General multi-PD cases require iterative approaches to solve the covariance optimization problem.

B. Multiple EDs

Multiple-ED VLC security studies use beamforming, artificial noise, signaling design, and deployment optimization to protect confidential transmissions. Results address diverse receiver configurations, channel knowledge assumptions, and ED locations.

  • MIMO systems with multiple EDs use BER minimization in protected zones and maximization elsewhere, or narrow-beam angle-diversity transmitters to reduce information leakage.
  • Hexagonal optical-network deployment provides the best secure-communication performance, whereas Poisson point process deployment provides the worst.
  • Single ED: For single-AU systems, studies use continuous and discrete signaling, including beamforming, zero-forcing, robust beamforming, and artificial-noise precoding.
  • Single ED: When ED channel state information is unavailable, transmission aligns the information signal with the AU and places jamming in the AU-orthogonal subspace.
  • Single ED: When the ED is close to the AU, the optimal information-signal amplitude fraction decreases, allocating more amplitude and power to jamming.
  • Multiple EDs: Multiple-ED systems employ continuous or discrete signaling and beamforming, including LED selection, stochastic-geometry analysis, and experimental artificial-noise precoding.

B. Multiple AUs

Secure MU-MISO VLC broadcast research considers confidential messages among legitimate users and protection against external EDs. Proposed designs include zero-forcing, artificial noise, and fairness-oriented precoding, evaluated with secrecy metrics.

  • Some studies treat legitimate AUs as potential EDs for messages intended for other users, while others include external EDs alongside the AUs.
  • Confidential broadcast transmission requires each AU to decode only its intended message while remaining unaware of the others.
  • Achievable secrecy sum-rate increases as user fairness decreases.
  • Reported work applies zero-forcing precoding, artificial-noise schemes, and max-min fairness SINR optimization among AUs.
  • Secrecy performance is evaluated using secrecy outage probability and ergodic secrecy rate, with and without AP cooperation.

V. THE SISO VLC WIRETAP SYSTEM

The SISO VLC wiretap system transmits multiple confidential messages through superposition coding under an amplitude constraint, with decoding based on NOMA and successive interference cancellation. Its literature review covers single- and multiple-AU settings with multiple EDs.

  • A single-AP transmitter superposes N confidential messages into one scalar signal subject to an amplitude constraint.
  • Because messages are transmitted simultaneously and non-orthogonally, AUs use NOMA detection with power-domain multiplexing and successive interference cancellation.
  • The received signals depend on AU and ED channel gains together with additive white Gaussian noise samples.
  • The review considers secure SISO VLC systems for both single-AU and multiple-AU configurations in the presence of multiple EDs.

B. Single AU

Single-AU VLC secrecy depends on whether the eavesdropper channel is weaker and on the chosen input distribution. For degraded SISO channels, discrete finite-support inputs provide the strongest reported lower bound, while unknown eavesdropper CSI motivates stochastic-geometry analysis.

  • When ρA,1 ≤ ρE, the SISO VLC wiretap channel is not degraded and its secrecy capacity is Cs = 0.
  • When ρE < ρA,1, the SISO VLC wiretap channel is strictly degraded and its secrecy capacity is nonzero.
  • Achievable secrecy rates have been derived under both continuous and discrete input distributions, including uniform, truncated Gaussian, TGN, and finite-support discrete distributions.
  • 105 independent Monte-Carlo trials were used to compare the upper bound UB, numerical secrecy capacity Cs, and lower bounds against A^2.
  • The finite-support discrete-input lower bound R+ s,1 is best because discrete finite-support inputs are optimal for the degraded SISO VLC wiretap channel.
  • Without ED channel-state information, stochastic geometry models random ED locations and averages achievable secrecy rates over those locations.

C. Multiple AUs

Multiple-AU VLC security research considers broadcast transmission, relay assistance, and randomly located eavesdroppers. Proposed schemes include direct transmission, cooperative jamming, decode-and-forward, and amplify-and-forward strategies.

  • Secure MU-SISO VLC broadcast systems send two confidential messages to two AUs in the presence of an external ED.
  • Relay-aided transmission strategies include cooperative jamming, decode-and-forward, and amplify-and-forward.
  • For these transmission schemes, researchers derived achievable secrecy regions and studied performance with respect to coverage geometry.
  • Other MU-SISO models represent the number of EDs with a Poisson point process and distinguish non-colluding from colluding eavesdroppers.

VI. HYBRID VLC/RF SYSTEMS

Hybrid VLC/RF systems combine VLC’s localized high-rate coverage with RF’s broader mobility support, but their joint broadcast nature creates security challenges. The literature also identifies unresolved signaling, channel, and multiuser secrecy problems.

  • VLC light is confined to small indoor areas and can be blocked by people, causing severe received-SNR fluctuations.
  • Hybrid VLC/RF integration combines VLC’s high data rates in specific areas with RF’s greater coverage for mobility support.
  • LiFi provides networked optical communication with infrared uplink and visible-light downlink transmissions.
  • Security in hybrid VLC/RF networks requires attention because both communication components have broadcast characteristics.
  • Existing work includes downlink hybrid RF/VLC wiretap systems and decode-and-forward relaying with ZF beamforming for mitigating eavesdropping.
  • Open problems include optimal signaling under VLC amplitude constraints and secrecy analysis for MU-MIMO VLC channels with colluding EDs and imperfect CSI.

C. Incorporating Realistic and Measurements Based VLC Channel Models

The paper identifies unrealistic assumptions in existing secure VLC channel models and highlights measurement-based models that incorporate mobility, orientation, and blockage for more practical secrecy analysis.

  • Current channel-model limitations: Existing secure VLC studies generally assume stationary or uniformly distributed receivers with fixed upward orientation.These assumptions omit mobility, device orientation, and link blockage.
  • Mobility and orientation: User mobility randomizes distance, transmission angle θ, and incidence angle ψ, while device orientation randomizes incidence angle ψ.The channel-gain formulation therefore depends on both movement and orientation effects.
  • Measurement-based models: Measurement-based models show that the incidence angle ψ follows a truncated Laplace distribution for stationary receivers and a truncated Gaussian distribution for mobile receivers.Measurements covered sitting and walking activities in portrait and landscape device modes.
  • Performance impact: Random device orientation can influence VLC user performance, with the effect depending on the user’s location.The cited work analytically derived SNR and BER distributions using a Laplace model.
  • Future research: Re-investigating secrecy performance with measurement-based, mobility, and orientation-aware models is proposed as a future direction for practical VLC PLS.The orientation-based random waypoint model incorporates both device orientation and user mobility.
  • Link blockage: Link blockage from human bodies or similar objects can interrupt communication, and extra transmit power cannot compensate for the resulting data loss.Multi-directional and omnidirectional receiver designs are discussed as blockage-mitigation approaches.

D. Security for Outdoor VLC systems

Outdoor VLC security remains an open area, especially for V2X and B2B communications, where realistic system configurations, interference, illumination, and weather effects require further investigation.

  • Outdoor VLC context: Outdoor VLC is less explored than indoor VLC because dual-use LEDs are less practical outdoors and interference and noise levels are higher.These conditions motivate dedicated outdoor channel and security analysis.
  • V2X communication: V2X communication covers vehicle-to-building, vehicle-to-infrastructure, vehicle-to-network, vehicle-to-pedestrian, and vehicle-to-vehicle links.It connects transportation elements including pedestrians, vehicles, roads, and cloud environments.
  • V2X communication: V2X communication is associated with improving traffic efficiency, saving resources, reducing accidents, and improving traffic management.The passage describes these as consequences of linking transportation elements through V2X.
  • B2B communication: VLC-based B2B communication can connect nearby buildings such as campuses, bank buildings, and headquarters to provide access to information, data, and media.B2B technologies may use RF or FSO communications, including visible-light and infrared spectra.
  • Security challenges: Security issues arise for VLC-based V2X and B2B communications, including concerns about large-scale system failures or malicious attacks in wireless vehicular networks.The paper identifies improving their security from a PLS perspective as future work.
  • Open research directions: Future outdoor VLC studies should derive measurement-based channel models incorporating interference and examine sunlight, nighttime illumination, and weather conditions.Careful system design is required to minimize impairment effects, although solar irradiance does not prevent high-speed VLC communication.
  • Survey scope: The survey reviews PLS across SISO, MISO, MIMO, and hybrid RF/VLC systems while identifying open problems for realistic VLC security techniques.It considers signaling, network geometry, legitimate receivers, eavesdroppers, and transmitter CSI availability.
Loading 1905.11450v1…