We investigate the fading cognitive multiple-access wiretap channel (CMAC-WT), in which two secondary-user transmitters (STs) send secure messages to a secondary-user receiver (SR) in the presence of an eavesdropper and subject to interference threshold constraints at multiple primary-user receivers (PRs). We design linear precoders to maximize the average secrecy sum rate for a multiple-input-multiple-output (MIMO) fading CMAC-WT under finite-alphabet inputs and statistical channel state information at STs. For this nondeterministic polynomial-time NP-hard problem, we utilize an accurate approximation of the average secrecy sum rate to reduce the computational complexity and then present a two-layer algorithm by embedding the convex-concave procedure into an outer-approximation framework. The idea behind this algorithm is to reformulate the approximated average secrecy sum rate as a difference of convex functions and then generate a sequence of simpler relaxed sets to approach the nonconvex feasible set. Subsequently, we maximize the approximated average secrecy sum rate over the sequence of relaxed sets by using the convex-concave procedure. Numerical results indicate that our proposed precoding algorithm is superior to the conventional Gaussian precoding method in the medium and high signal-to-noise ratio (SNR) regimes.
We investigate the fading cognitive multiple-access wiretap channel (CMAC-WT), in which two secondary-user transmitters (STs) send secure messages to a secondary-user receiver (SR) in the presence of an eavesdropper and subject to interference threshold constraints at multiple primary-user receivers (PRs). We design linear precoders to maximize the average secrecy sum rate for a multiple-input-multiple-output (MIMO) fading CMAC-WT under finite-alphabet inputs and statistical channel state information at STs. For this nondeterministic polynomial-time NP-hard problem, we utilize an accurate approximation of the average secrecy sum rate to reduce the computational complexity and then present a two-layer algorithm by embedding the convex-concave procedure into an outer-approximation framework. The idea behind this algorithm is to reformulate the approximated average secrecy sum rate as a difference of convex functions and then generate a sequence of simpler relaxed sets to approach the nonconvex feasible set. Subsequently, we maximize the approximated average secrecy sum rate over the sequence of relaxed sets by using the convex-concave procedure. Numerical results indicate that our proposed precoding algorithm is superior to the conventional Gaussian precoding method in the medium and high signal-to-noise ratio (SNR) regimes.
In this paper, we investigate the secrecy rate of finite-alphabet communications over multiple-input-multiple-output-multiple-antenna eavesdropper (MIMOME) channels. Traditional precoding designs based on Gaussian input assumption may lead to substantial secrecy rate loss when the Gaussian input is replaced by practical finite-alphabet input. To address this issue, we investigate linear precoding designs to directly maximize the secrecy rate for MIMOME systems under the constraint of finite-alphabet input. By exploiting the theory of Karush-Kuhn-Tucker (KKT) analysis and matrix calculus, we first present necessary conditions of the optimal precoding design when instantaneous channel-state information (CSI) of the eavesdropper is known at the transmitter. In this light, an iterative algorithm for finding the optimal precoding matrix is developed, utilizing a gradient decent method with backtracking line search. Moreover, we find that the beamforming design in MIMONE systems, which is a secrecy-capacity-achieving approach for Gaussian signaling, no longer provides the maximum secrecy rate for finite-alphabet input data. This case is substantially different from the Gaussian input case. In addition, we derive the closed-form results on the precoding matrix, which maximizes the secrecy rate in the low signal-to-noise ratio (SNR) region, and reveal the optimal precoding structure in the high-SNR region. A novel jamming signal generation method that draws on the CSI of the eavesdropper to additionally increase the secrecy rate is further proposed. The precoding design with only statistical CSI of the eavesdropper available at the transmitter is also considered. Numerical results show that the proposed designs provide significant gains over recent precoding designs through a power control policy and the precoding design with the Gaussian input assumption in various scenarios.
This paper studies multiple-input multiple-output multiple access wiretap channels (MAC-WT) where an eavesdropper tries to tap the communication between multiple legitimate transmitters and a legitimate receiver. In this system, we propose precoder optimization methods at the transmitters in order to maximize the sum secrecy rate performance. Although this problem can be solved by the well-known difference of convex (DC) programming, we present a more efficient algorithm whose computational complexity is much lower than that of the conventional DC approach. By investigating the Karush-Kuhn-Tucker conditions, it is confirmed that the proposed low-complexity algorithm achieves the same performance as the conventional DC method. Our analysis also reveals that the proposed algorithm ensures global optimality for multiple-input single-output MAC-WT cases, while binary power control is optimal for single-input multiple-output scenarios. Simulation results demonstrate the efficacy of the proposed precoding methods.
In this paper, we investigate the secrecy rate of finite alphabet communications over multiple-input, multiple-output, multiple-antenna eavesdropper (MIMOME) systems. Traditional precoder designs at the transmitter for achieving secrecy capacity (maximum secrecy rate) for MIMOME systems are developed according to the assumption of Gaussian input signals. Such designs may risk substantial secrecy rate loss when Gaussian inputs are replaced by practical finite alphabet inputs. To address this issue, we propose a linear precoding design to directly maximize the secrecy rate for MIMOME systems under the constraint of finite alphabet input. Exploiting convex optimization and matrix calculus, we present necessary conditions required of the optimal precoding design and develop an iterative algorithm for finding an efficient precoder. With finite alphabet input signals, maximum transmission power no longer corresponds to maximum secrecy rate as in the case of Gaussian input. We further derive closed-form results on the optimal transmission design for maximizing secrecy rate in low signal-to-noise ratio (SNR) region and near-optimal transmission power in a high SNR region.
In this letter, we study the secrecy precoding problem for simultaneous wireless information and power transfer (SWIPT) in a cognitive multiple-input multiple-output (MIMO) broadcast channel. We adopt an artificial noise (AN)-aided precoding scheme and formulate the problem as a secrecy rate maximization (SRM) problem, which is subject to both an interference power constraint imposed to protect the primary user (PU) and an energy harvesting constraint required by the secondary energy receiver (SER). Since the formulated SRM problem constitutes a difference convex (DC)-type programming problem, we solve it by employing a successive convex approximation (SCA) method. With the SCA method, the nonconvex part of the SRM problem can be locally linearized to its first-order Taylor expansion. Then, relying on solving a series of convexified optimization problems, an iterative precoding algorithm is developed. Numerical simulations are also provided to demonstrate the proposed algorithm. Results show that our algorithm can achieve a near-optimal performance with guaranteed convergence.
In this paper, we investigate the precoding scheme for multiple-input multiple-output multiple antennas eavesdropper (MIMOME) channels under the constraint of finite-alphabet inputs to ensure secure transmission between the source and the destination in the presence of an eavesdropper. With finite-alphabet inputs, there are no closed-form solutions for mutual information (MI) and its gradient-minimum mean-square error (MMSE) matrix. Therefore traditional precoding schemes based on direct calculation of MI and MMSE matrix through numerical integration involve prohibitive computational burden. To address this problem, this paper first derives asymptotically optimal lower bounds of MI and ergodic MI with which the computational complexity can be significantly reduced compared with the direct computation of MI. With the derived lower bounds, a low-complexity precoding scheme is proposed to maximize the secrecy rate, wherein a gradient descent based algorithm is used to obtain the precoding result. In the low and high signal-to-noise ratio (SNR) regions, we demonstrate that the proposed precoding scheme gives the same result as directly maximizing the secrecy rate, which indicates that the proposed low-complexity precoding scheme is asymptotically optimal.
This paper studies a cognitive multiple-input multiple-output (MIMO) network, in which a secondary transmitter aims to send confidential messages to its receiver in the presence of one multi-antenna eavesdropper, while has the interference temperature limit (ITL) for the primary user. Assuming that all the channels are imperfect and the errors are modeled by deterministically bounded model, we consider robust secure beamforming (RSB) design problem and our goal is to maximize the worst-case achievable secrecy rate subject to the total transmit power constraint and the ITL for the primary user. The formulated optimization problem is non-convex while the channel uncertainties make it more difficult to cope with. To address it, an inexact block coordinate descent (IBCD) approach is proposed to obtain the near-optimal solution, where we firstly linearize the two log-det terms of the objective function and then, the semi-infinite constraints with the bounded channel uncertainties are reformulated as linear matrix inequalities (LMIs). Finally, numerical results validate the performance of our proposed RSB algorithm.