False data injection (FDI) attacks, as a new class of cyberattacks, bring a sever threat to the security and reliable operation of smart grid by damaging the state estimation of the power system. To address this issue, an extreme learning machine (ELM)-based One-Class-One-Network (OCON) framework is proposed for detecting the FDI attacks in this work. Under this framework, to effectively detect bus-based FDI attacks and identify the bus node being attacked, the subnets of state identification layer in OCON adopt the ELM algorithm to accurately divide the false data and the normal data. After that, a global layer is employed to analyze whether the bus node associated with its corresponding subnet is attacked by false data utilizing the results from the state identification layer. Finally, in order to improve the resilience of the power system, a prediction recovery strategy is proposed to remedy the detected false data by exploiting the spatial correlation of power data. The proposed framework is tested on IEEE 14 bus system using real load data from New York independent system operator (NYISO). Simulation results demonstrate that the proposed framework not only accurately recognizes the multiple bus nodes under the FDI attacks, but also efficiently recovers the data injected by false data.

With the development of artificial intelligence (AI), wireless channel prediction based on deep learning (DL) has become a hot research issue. Channel prediction plays an important role in channel state information (CSI) feedback enhancement in AI-native air interface. To better predict the CSI, this paper investigates the Transformer network based channel prediction. Firstly, real channel data are obtained in Beijing-Tianjin railway line, and the channel prediction datasets are constructed through preprocessing. After formulating the channel prediction problem, a channel prediction model based on the Transformer network is newly proposed. The unique multi-head attention mechanism and position encoding of the Transformer network enable the proposed model to have more powerful parallel computation capability and better global information capture capability. Then, the hyper-parameters of the model are determined by autocorrelation analysis and cross-validation. Finally, the performance of the proposed model is evaluated in terms of prediction accuracy and space and time computational complexity using several evaluation metrics, and is compared with classical DL models. It is shown that the proposed model possesses higher prediction performance in the appropriate range of computational complexity.

For massive multiuser multiple input multiple output (MIMO) (MU-MIMO) systems, pilot contamination (PC) and data interference (DI) are two main factors that affect the accuracy of the superimposed pilot (SP)-based channel estimation, provided that the channel coherence time is less than the total number of users. In this paper, to tackle the effect of PC, we propose a block-diagonal Grassmannian Line Packing (BD-GLP) approach, in which the specific sequences for different cells are first designed based on GLP, and then are block-diagonally extended to form the SP matrices for the users in different cells. To also alleviate the effect of DI on the channel estimation, with the designed SP matrices, an iterative channel estimation (ICE) method based on Tikhonov regularization (TR) is presented. The impact that the proposed ICE method has on the signal to interference plus noise ratio (SINR) with matched filtering (MF) receiver is theoretically analyzed. Moreover, the optimal power allocation factor to SP by maximizing the spectral efficiency (SE) of the target cell is also derived. Numerical results show that the proposed ICE method with the designed SP matrices effectively mitigates the effects of the PC and DI, and hence improves the accuracy of channel estimation and further the SE of the massive MU-MIMO systems.

Interference is believed to be the most significant bottleneck for the next-generation wireless networks to achieve high throughout. Interference alignment (IA), as a novel interference management scheme to break through the traditional interference cancellation, not only makes the complete mitigation of interference possible, but also achieves a theoretical breakthrough in promoting the wireless network capacity region. In this paper, by combining the space and time, we proposed a linear space-time IA (LST-IA) algorithm based on the extension of the channel in time dimension for K-user multi-input multi-output (MIMO) interference channel (IC). The proposed LST-IA scheme effectively reduces the number of antennas required for eliminating interference completely in systems, and the closed-form solution of precoding matrices and detector matrices are obtained as well. Compared with the classical IA algorithms, the simulation results demonstrate that the proposed scheme shows distinguished advantages in terms of sum-rate and bit error rate (BER) in the strong interference communication scenarios.

In order to reduce the implementation complexity and hardware cost of millimeter wave (mmWave) massive multiple input multiple output (mMIMO) systems, we propose a dynamically-connected hybrid precoding (DC-HP) scheme in this paper. In the scheme, to exploit the subchannels difference of the users in the system, a dynamic connection network and an analog precoding matrix are first jointly designed. After that, the matched-filter (MF) or zero-forcing (ZF) criterion is used to design the digital precoding matrix with the equivalent channel after analog precoding. Simulation results show that the DCHP scheme achieves a better performance than the traditional sub-connected hybrid precoding (SC-HP) and adaptive hybrid precoding (AHP) schemes, while it also enjoys a much lower hardware implementation complexity and hardware power consumption than the full digital baseband precoding (FD-BP) and fully-connected hybrid precoding (FC-HP) schemes.