We investigate in this paper the achievable secrecy rate of the multiple-input single-output (MISO) visible light communication (VLC) Gaussian wiretap channel with single user and single eavesdropper. We consider the cases when the location of eavesdropper is known or unknown to the transmitter. In the former case, we derive the optimal beamforming in closed form, subject to constrained inputs. In the latter case, we apply robust beamforming. Furthermore, we study the achievable secrecy rate when the input follows the truncated generalized normal (TGN) distribution. We present several examples which demonstrate the substantial improvements in the secrecy rates achieved by the proposed techniques.
We investigate in this paper the achievable secrecy rate of the multiple-input single-output (MISO) visible light communication (VLC) Gaussian wiretap channel with single user and single eavesdropper. We consider the cases when the location of eavesdropper is known or unknown to the transmitter. In the former case, we derive the optimal beamforming in closed form, subject to constrained inputs. In the latter case, we apply robust beamforming. Furthermore, we study the achievable secrecy rate when the input follows the truncated generalized normal (TGN) distribution. We present several examples which demonstrate the substantial improvements in the secrecy rates achieved by the proposed techniques.
We study the secrecy capacity of the multiple- input single-output (MISO) Gaussian wiretap visible light communication (VLC) channel. We study a typical VLC scenario with one transmitter, one legitimate receiver, and one eavesdropper. Specifically, we compute the achievable secrecy rate for various input signaling distributions, including the truncated generalized normal (TGN) and uniform distributions. The transmitter is equipped with multiple light sources, while the legitimate and unauthorized receivers are each equipped with a single photodetector. We analyze the achievable secrecy rates via transmit beamforming and artificial noise. In addition, both zero-forcing beamforming and robust beamforming are considered. In the former case, the location of the eavesdropper is assumed to be known, whereas in the latter case, the location of the eavesdropper is unknown. Our numerical results show that the secrecy rate achieved by the TGN distribution is significantly improved as compared to those achieved by the truncated Gaussian and uniform distributions, for both zero-forcing beamforming and robust beamforming. We also derive an upper bound on the achievable secrecy capacity that we used to assess the closeness of the achievable secrecy rates to the derived bound.
We study in this paper the achievable secrecy rate of the Gaussian multiple-input single-output (MISO) visible light communication (VLC) in the presence of randomly located eavesdroppers. We consider a system model comprising a transmitter (Alice) equipped with multiple fixtures of LEDs, one legitimate receiver (Bob) equipped with a single photo-diode (PD) and a group of randomly located eavesdroppers, each equipped with a single PD. The channel is modeled as deterministic, real- valued and subject to an amplitude constraint, through which Alice wants to communicate privately with Bob. We consider the case where the eavesdroppers are colluding together, i.e., they act jointly in eavesdropping on the communication between Alice and Bob. We adopt the truncated generalized normal (TGN) as input signaling and beamforming as transmission strategy. At first, we derive a closed-form expression of the average achievable secrecy rate as a function of the density of eavesdroppers using stochastic geometry. Then, we investigate the optimal beamformer that maximizes the average achievable secrecy rate of the system. Finally, the analysis is verified by Monte Carlo simulations.
This paper investigates the secrecy performance of the multiple-input single-output visible light communication (VLC) wiretap channel. The considered system model comprises three nodes: a transmitter (Alice) equipped with multiple fixtures of LEDs, a legitimate receiver (Bob), and an eavesdropper (Eve), each equipped with one photo-diode. The VLC channel is modeled as a real-valued amplitude-constrained Gaussian channel. Eve is assumed to be randomly located in the same area as Bob. Due to this, artificial noise-based beamforming is adopted as a transmission strategy in order to degrade Eve's signal-to-noise ratio. Assuming discrete input signaling, we derive an achievable secrecy rate in a closed-form expression as a function of the beamforming vectors and the input distribution. We investigate the average secrecy performance of the system using stochastic geometry to account for the location randomness of Eve. We also adopt the truncated discrete generalized normal (TDGN) as a discrete input distribution. We present several examples through which we confirm the accuracy of the analytical results via Monte Carlo simulations. The results also demonstrate that the TDGN distribution, albeit being not optimal, yields performance close to the secrecy capacity.
We investigate in this paper the achievable secrecy sum-rate of the multiple-input single-output (MISO) visible light communication (VLC) broadcast Gaussian channel with confidential messages. We consider a system model comprising (K +1) nodes: a transmitter (Alice) equipped with N fixtures of LEDs and K (K ≤ N) spatially dispersed users, each equipped with a single photo-diode (PD). The channel is modeled as deterministic and real-valued, subject to amplitude and power constraints, through which Alice wants to transmit in one channel use K confidential messages. We propose a new precoding scheme based on the eigenvalues of the pencils of the different K MISO VLC channels in order to maximize the achievable secrecy sum-rate of the overall system. The motivation behind this scheme is based on the fact that the optimal precoding scheme maximizing the achievable secrecy rate of the single user MISO VLC wiretap channel is expressed through the eigenvector associated to the largest eigenvalue of the wiretap channel's pencil. We investigate both cases, namely, when the locations of users are known, i.e., perfect channel state information (CSI) is available to Alice, or unknown, i.e., imperfect CSI. Furthermore, we adopt the truncated generalized normal (TGN) distribution as input signaling. We present several examples which demonstrate the substantial improvements in the secrecy sum-rates achieved by the proposed techniques compared to those achieved by zero-forcing (ZF) precoding.
This paper investigates a multiple-input multiple-output (MIMO) visible light communication (VLC) wiretap channel consisting of a transmitter, a legitimate receiver, and an eavesdropper. The optical input is subject to both peak- and average-intensity constraints. By applying the generalized entropy-power inequality to truncated exponential inputs, we derive a novel closed-form expression for the achievable secrecy rate for general MIMO-VLC configurations. To enhance transmission confidentiality, a fully-connected beamforming scheme is proposed, along with a low-complexity sub-connected alternative. Although the resulting beamforming design problems are non-convex, they are efficiently addressed by transforming them into a sequence of convex subproblems solvable via the successive convex approximation framework. Numerical results demonstrate that the proposed schemes achieve significant secrecy performance improvements compared with the benchmark scheme.
We study in this paper the secrecy performance of the multiple-input single-output (MISO) visible light communication (VLC) channel in the presence of randomly located eavesdroppers. We consider a system model comprising a transmitter, equipped with multiple fixtures of LEDs, one legitimate receiver equipped with a single photo-diode (PD) and a group of randomly located (and colluding) eavesdroppers, each equipped with a single PD. We derive a closed-form expression for the average achievable secrecy rate as a function of the beamforming vector, the input distribution, and the location density of the eavesdroppers using stochastic geometry. We then investigate the optimal beamformer that maximizes the average achievable secrecy rate. We corroborate the analytical results through Monte Carlo simulations and we demonstrate substantial improvements provided by the proposed scheme over existing ones.