Utilizing acoustic signals from the rotor or power engines is an important approach of identifying and localizing the drone. When localize the near-field drone, the uncertainties in the environment may introduce errors both in the signals propagation speed and the sensors positions. To address the issue, this letter proposes a double null space projection (DNSP) closed-form algorithm based on time difference of arrival (TDOA) and frequency difference of arrival (FDOA) measurements to achieve high precision motion parameter estimation of the drone. Firstly, the null space projection method is introduced to the constructed localization equation to obtain the initial solution of drone motion parameter. However, due to the relationship between the actual unknowns and nuisance parameters in the projection process is neglected, this solution is suboptimal. Secondly, DNSP ultimately refines the localization performance by error compensation. Simulation results and performance analysis show that DNSP has Cramér–Rao lower bound (CRLB) localization accuracy under small noise.
2 April 2024 in IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
It is well known that distillation learning has the ability to enhance the performance of a light (student) model by transferring knowledge from a heavy (teacher) model, without incurring additional computational and storage costs. This article proposes an improved decoupled knowledge distillation (DKD) strategy for hyperspectral image (HSI) classification. A spatial feature blurring (SFB) module is designed to improve the classification performance of the student network when using DKD strategy. The SFB module utilizes randomly initialized two-dimensional standard normal distribution tensors to blur the spatial features of HSI, which increases the complexity of the data. This aligns with the characteristics of DKD, which transfers more useful knowledge under the condition of sample complexity. To effectively transfer knowledge, this article proposes a robust teacher network named the dual-branch spatial transformer-spectral transformer (DBSTST) network. This network describes the spatial and spectral long-range dependencies of HSI, addressing the limitations of convolutional neural networks (CNNs) in capturing only local features due to their fixed receptive fields. More specifically, the DBSTST network adopts spatial transformer-spectral transformer (STST) which is composed of a parallel spatial-spectral multi-head self-attention (PS2MHSA) module, aiming to describe pixel-level spatial long-range dependencies and spectral correlations in HSI. Simultaneously, the introduction of spatial-spectral positional embedding into PS2MHSA enhances positional awareness. We demonstrated the effectiveness of our proposed method on four publicly available HSI datasets. The student network achieves classification performance improvement and surpasses some other networks. Moreover, when compared to state-of-the-art classification methods, the DBSTST network also exhibits significant improvements in classification performance.
13 November 2023 in IEEE Transactions on Cognitive Communications and Networking
Covert communication aims to enhance the security of the wireless communication by hiding the transmission behavior from the eavesdropper. In this paper, a spectrum mask-based multi-antenna covert communication strategy is proposed in Internet of Things (IoT) networks. Specifically, the transmitter superimposes its messages on the existing overt channel to prevent the transmission being detected by the adversary eavesdropper, which can significantly enhance communication covertness without the need for additional spectrum resources. To eliminate the interference cause to the existing overt channel, multi-antenna technology and improper Gaussian signaling (IGS) are leveraged at the covert transmitter. Moreover, maximum ratio transmission (MRT) is adopted at the covert transmitter can fully exploit the additional degrees of freedom (DoF) that the multi-antenna provides to achieve reliable and covert transmission. Firstly, the optimal detection threshold at the eavesdropper is analyzed, and minimum error detection probability is then calculated, which indicates the worst-case from the covert transmitter’s perspective. In this scenario, a joint transmit power and IGS factor optimization problem is proposed to maximize the covert throughput under the covertness constraints. Secondly, the optimization problem is divided into three subproblems based on the power relationship, then the optimal transmit power and IGS factor are solved, respectively. Finally, extensive simulation is executed to demonstrate that under the given covertness constraints, the proposed strategy can effectively improve the covert communication performance of the multi-antenna system in IoT networks.
This paper proposes a novel strategy to enhance covert communication performance through a full-duplex communication system assisted by Unmanned Aerial Vehicles (UAV) and Intelligent Reflecting Surfaces (IRS). In the proposed system, the base station (Alice) leverages the IRS mounted on UAV to reflect covert information towards the legitimate full-duplex receiver (David). As a full-duplex receiver, David actively generates dynamic artificial noise (AN) to deceive the warden (Eve) and obscure her detection capability. Notably, due to the adversarial nature of the scenario, Eve’s exact location remains unknown to the legitimate parties. Within this framework, we first derive the minimum detection error probability (DEP) for Eve under the worst-case scenario, where Eve employs the optimal detection threshold. Subsequently, we formulate an optimization problem aimed at maximizing the covert transmission rate while satisfying the covertness constraint by alternately optimizing the IRS phase shifts and the three-dimensional (3D) trajectory of UAV. Simulation results validate the effectiveness of the proposed strategy in enhancing covert communication performance, particularly in non-line-of-sight (NLOS) environments.
A robust algorithm for locating the signal source is proposed based on information fusion from different domains to improve the passive localization accuracy under the nonline-of-sight (NLoS) environment. The localization scenario is established by utilizing the information about energy domain, time domain, and frequency domain. The Cramer–Rao lower bounds are derived and the performance improvement is proved in different domains. By substituting NLoS bias with the statistical average, the localization equations of information fusion are established. A nonconvex localization problem is transformed into a generalized trust domain subproblem by introducing the squared range and the weighted least squares. The position and velocity of the source are obtained by using the bisection procedure. An iterative method is used to estimate the NLoS deviation and refine the position accuracy simultaneously. Simulation results demonstrate the efficiency and robustness of the proposed algorithm.
The finite battery capacity, limited spectrum resource and the transmission security are three challenges in designing mobile edge computing (MEC) systems. In this paper, a framework for wireless powered secure multiple-input single-output (MISO) cognitive-based MEC-enabled networks is proposed, which integrates several technologies: physical-layer security, cooperative relaying, cognitive radio (CR), wireless power transfer (WPT) and MEC. Two optimization problems are formulated to maximize the number of computation bits (NCB) of the mobile device (MD) and minimize the total transmission power (TTP) of the primary transmitter (PT) and the SBS, respectively. The formulated problems are non-convex and hard to solve. Two two-phase methods with block coordinate decent (BCD) and Lagrangian dual decomposition methods are proposed to jointly design the beamforming vector of the PT, beamforming matrix of the secondary base station (SBS), the central processing unit (CPU) frequency, the transmit power, the number of the offloaded bits and the offloading time of the MD. Simulation results are presented to show the effectiveness of the proposed methods.