In recent years, wireless federated learning (FL) has been proposed to support the mobile intelligent applications over the wireless network, which protects the data privacy and security by exchanging the parameter between mobile devices and the base station (BS). However, the learning latency increases with the neural network scale due to the limited local computing power and communication bandwidth. To tackle this issue, we introduce model pruning for wireless FL to reduce the neural network scale. Device selection is also considered to further improve the learning performance. By removing the stragglers with low computing power or bad channel condition, the model aggregation loss caused by model pruning can be alleviated and the communication overhead can be effectively reduced. We analyze the convergence rate and learning latency of the proposed model pruning method and formulate an optimization problem to maximize the convergence rate under the given learning latency budget via jointly optimizing the pruning ratio, device selection, and wireless resource allocation. By solving the problem, the closed-form solutions of pruning ratio and wireless resource allocation are derived and the threshold-based device selection strategy is developed. Finally, extensive experiments are carried out to demonstrate that the proposed model pruning algorithm outperforms other existing schemes.

The escalating crosstalk and interference in wireless services necessitate effective techniques like single-channel blind source separation (SCBSS). Deep Learning (DL) has shown promise in enhancing spectrum efficiency in wireless communications. Existing DL-SCBSS methods primarily focus on speech separation and speech enhancement, limiting their direct application to communication signals due to waveform and time scale differences. To overcome this, we introduce a complex time-domain dilated convolutional recurrent network (CTDCRN) featuring a complex hierarchical encoder (CHE) for complex signal representation. CTDCRN employs a convolutional recurrent network with complex dilated convolution module (CDCM) and complex LSTM (CLSTM) to extract precise features from communication signals. CDCM models short and long-term dependencies using depth-wise dilated convolutions, enhancing individual signal separation. Unlike speech separation networks, CTDCRN processes in-phase and quadrature components (I/Q) of complex signals. Simulation results showcase CTDCRN’s superiority over traditional methods and real-valued networks in Pearson’s correlation coefficient and bit error rate (BER).

In this paper, a multi-cell multiple-input multiple-output (MIMO) uplink system is considered. For uplink transmission, to implement decoding, the base station (BS) generally needs to estimate channels with the aid of training sequences. Obviously, for full frequency reuse systems, training also suffers from inter-cell interference, namely pilot contamination. In this paper, the impact of pilot contamination on the performance of the considered system is studied based on the maximum likelihood (ML) decoder. Specifically, an exact analytical expression of pair-wise error probability (PEP) is derived, and then, a lower bound and an upper bound of PEP are given, to explicitly show the impact of pilot contamination. With the detailed analysis of these bounds, we can discover: (1) pilot contamination can result in error floor, provided that the number of BSs in the considered system is larger than the length of a frame; and (2) if the length of a frame is larger than or equal to the number of BSs, and an appropriate coding scheme is applied, error floor can be removed. Finally, a coding design criterion is proposed. Based on this criterion, it can be shown that the considered system can achieve the same diversity to what can be achieved by a single isolated cell system.

Vein contraction and venous compression typically caused by low temperature and excessive placement pressure can blur the captured finger vein images and severely impair the quality of extracted features. To improve the quality of captured finger vein image, this paper proposes a 26-layer generator network constrained by Neighbors-based Binary Patterns (NBP) texture loss to recover the clear image (guessing the original clear image). Firstly, by analyzing various types and degrees of blurred finger vein images captured in real application scenarios, a method to mathematically model the local and global blurriness using a pair of defocused and mean blur kernels is proposed. By iteratively and alternatively convoluting clear images with both kernels in a multi-scale window, a polymorphic blur training set is constructed for network training. Then, NBP texture loss is used for training the generator to enhance the deblurring ability of the network on images. Lastly, a novel network structure is proposed to retain more vein texture feature information, and two residual connections are added on both sides of the residual module of the 26-layer generator network to prevent degradation and overfitting. Theoretical analysis and simulation results show that the proposed neighbors-based binary-GAN (NB-GAN) can achieve better deblurring performance than the the-state-of-the-art approaches.

A massive multiple-input multiple-output (MIMO) system, which utilizes a large number of antennas at base stations to communicate with multiple user terminals each with a single antenna, is one of the most promising techniques for future wireless communications systems. A successful massive MIMO implementation relies on accurate channel estimation, which is typically performed through pilot sequences. However, the channel estimation performance or massive MIMO performance is limited by pilot contamination due to unavoidable reuse of pilot sequences from terminals in neighboring cells. In this paper, a semi-blind decoding method based on independent component analysis (ICA), channel energy levels, and reference bits is proposed. Specifically, the proposed decoding method uses ICA to separate and decode the received signals and to estimate the channels. The estimated channel energy is used to differentiate the in-cell signals and the neighboring cell signals, and reference bits are applied to identify a desired signal among signals within a cell. The analytical performance results of the proposed decoding method are derived. Numerical results show that the proposed semi-blind decoding method has better bit error rate performance and higher transmit efficiency than the traditional minimum mean-square error decoding method and the singular value decomposition-based decoding method.