Deep learning based extra-large scale multiple-input multiple-output (XL-MIMO) channel estimation faces two critical challenges: the strong reliance on labeled channel data and the high computational complexity. To address these limitations, this paper proposes a knowledge distillation transformer channel estimation (KDT-CE) framework that enables the true label-free training of a lightweight student model distilled from a pretrained teacher model. The student model combines the convolutional and Transformer-based structures, achieving a decent trade-off between the estimation accuracy and computational cost. In the training stage, a dual-path knowledge distillation mechanism is developed, which transfers the high-quality pseudo labels and latent channel features extracted from the teacher model to the student model for a better distillation efficiency. Simulation results show that the student model trained by KDT-CE outperforms existing self-supervised baseline schemes and approaches the teacher model with the number of parameters and computational complexity respectively reduced by about 20× and 6×, especially at low signal-to-noise ratio conditions.

deep learning (DL) has been widely applied to enhance automatic modulation classification (AMC). However, the elaborate AMC neural networks are susceptible to various adversarial attacks, which are challenging to handle due to the generalization capability and computational cost. In this article, an explainable DL based defense scheme, called Shapley additive explanation enhanced adversarial fine-tuning (SHAP-AFT), is developed in the perspective of disclosing the attacking impact on the AMC network. By introducing the concept of cognitive negative information, the motivation of using SHAP for defense is theoretically analyzed first. The proposed scheme includes three stages, i.e., the attack detection, the information importance evaluation, and the AFT. The first stage indicates the existence of the attack. The second stage evaluates contributions of the received data and removes those data positions with negative Shapley values corresponding to the dominating negative information caused by the attack. Then the AMC network is fine-tuned based on adversarial adaptation samples using the refined data pattern. Simulation results show the effectiveness of the Shapley value as the key indicator as well as the superior defense performance of the proposed SHAP-enhanced adversarial fine-tuning (SHAP-AFT) scheme in face of different attack types and intensities.

In hierarchical cognitive radio networks, the edge or cloud servers utilize the data collected by the edge devices for modulation classification, which, however, is faced with problems of the computation load, transmission overhead, and data privacy. In this article, an edge learning (EL)-based framework jointly mobilizing the edge device and the edge server for intelligent co-inference is proposed to realize the collaborative automatic modulation classification (C-AMC) between them. A spectrum semantic compression neural network is designed for the edge device to compress the collected raw data into a compact semantic embedding that is then sent to the edge server via the wireless channel. On the edge server side, a modulation classification neural network combining the bidirectional long-short term memory and attention structures is elaborated to determine the modulation type from the noisy semantic embedding. The C-AMC framework decently balances the computation resources of both the sides while avoiding the high transmission overhead and data privacy leakage. Both the offline and online training procedures of the C-AMC framework are elaborated. The compression strategy of the C-AMC framework is also developed to further facilitate the deployment, especially for the resource-constrained edge device. Simulation results show the superiority of the EL-based C-AMC framework in terms of the classification accuracy, computational complexity, and the data compression rate as well as reveal useful insights paving the practical implementation.

For millimeter wave (mmWave) systems with large-scale arrays, hybrid processing structure is usually used at both transmitters and receivers to reduce the complexity and cost, which poses a very challenging issue in channel estimation, especially at the low transmit signal-to-noise ratio regime. In this paper, deep convolutional neural network (CNN) is employed to perform wideband channel estimation for mmWave massive multiple-input multiple-output (MIMO) systems. In addition to exploiting spatial correlation, our joint channel estimation approach also exploits the frequency correlation, where the tentatively estimated channel matrices at multiple adjacent subcarriers are input into the CNN simultaneously. The complexity analysis and numerical results show that the proposed CNN based joint channel estimation outperforms the non-ideal minimum mean-squared error (MMSE) estimator with reduced complexity and achieves the performance close to the ideal MMSE estimator. It is also quite robust to different propagation scenarios.

In this paper, we propose a simple and efficient approach to reduce the overhead of downlink channel estimation and feedback using linear regression (LR) and support vector regression (SVR) in machine learning. Specifically, we divide the indexes of the antennas at the base station (BS) into two sets. We first use some well estimated channel samples to train a regression model, where the channel state information (CSI) corresponding to one set is used as input while the other is output. In the online channel estimation phase, only the CSI of the antennas in one set needs to be estimated and the CSI of the antennas in the other set can be predicted by inputting the estimated CSI into the trained regression model. Numerical results show the proposed approach can reduce the overhead of both downlink pilot and uplink feedback considerably and thus can improve the downlink achievable rate significantly compared with the existing schemes.

In this paper, a hybrid time-division duplex access scheme with confidence factor based weighted cooperative sensing for the femtocells has been proposed in cognitive massive multiple-input multiple- output (massive MIMO) system to improve the detection performance and achieve the high throughput while alleviate the influence to the macrocell users as little as possible. In the cognitive massive MIMO system, femtocell base stations (FBSs) sense the spectrum through cooperative scheme and decide whether to access the channel or not. Firstly, all the FBSs sense the macrocell users in the uplink channel and the macrocell BS in the downlink channel, and make dynamical clustering based on the channel conditions. After the information exchange within each cluster, the FBSs in each cluster update their confidence factors (CFs) and weight factors (WFs) according to the detection results. With consideration of the CFs and WFs, more accurate sensing results would be obtained. Simulation results and theoretical analysis show that the proposed scheme can get more accurate sensing results and better throughput tradeoff for the femtocells.