To mitigate the significant degradation in direction of arrival (DOA) estimation performance caused by time-varying axial deviation (TVAD) in acoustic vector sensor (AVS) array under non-uniform noise, a two-step least squares fitting (TSLSF) technique is presented in this paper. Firstly, a model for the AVS array incorporating TVAD is established by introducing axial deviation parameters into datasets from various subtime periods (STPs). Then, to treat the noise vectors of each channel in the AVS array as virtual sparse signals, a novel AVS array manifold matrix is formulated. After that, to estimate the TVAD matrix, sparse signals, and noise vector, two cost functions are constructed based on the principle of weighted least squares. Moreover, their analytical expressions were derived. Furthermore, to handle the effects of TVAD on DOA estimation performance over the entire observation period, the focusing technology is adopted to transform datasets with TVAD from different STPs into the desired dataset. Simulation experiments confirmed the effectiveness and resilience of the proposed method using an AVS array in conjunction with TVAD in the presence of non-uniform noise.

This paper considers device-to-device (D2D) communication underlaying cellular networks, where frequency resources are shared between D2D users and cellular users. When D2D users reuse the frequency resources occupied by the cellular users, interference could be produced among the two kinds of users. By means of the fractional frequency reuse approach, a traversal frequency reuse scheme is proposed, in which central D2D users in each cell reuse the frequency bands for neighbouring edge cellular users in sequence according to a counterclockwise direction. The proposed reuse scheme effectively minimizes the number of intra-cell interference links, leading to an improvement of the average sum rate. Subsequently, based on the proposed traversal reuse strategy, a deep Q-learning algorithm is implemented for power control. Simulation results demonstrate that the proposed power control algorithm outperforms other traditional methods in terms of sum rate.

As massive MU-MIMO systems scale, managing high computational demands and ensuring fairness among users become crucial challenges. Traditional scheduling methods often fall short in dynamically adapting to changing environments and balancing multi-dimensional performance metrics. To address these challenges, we enhance the advantage actor-critic (A2C) framework by integrating convolutional neural networks (CNNs) and Transformers for massive MU-MIMO systems. The CNN components specialize in extracting localized channel-state features, while the Transformers dynamically model inter-user dependencies through attention mechanisms. Specifically, we innovatively embed convolutional layers within the Transformer encoder and employ an auto-regressive decoder, reformulating the user group selection as a sequential decision-making process based on conditional probabilities. This represents the first application of a hybrid CNN-Transformer architecture for discrete scheduling decisions in MU-MIMO systems. To ensure balanced performance, we introduce a multi-metric reward function that incorporates multiple metrics rather than a single performance indicator. Our reward is calculated as the product of the selected user's spectral efficiency (SE) and the Jain fairness index (JRI) during scheduling. Simulations demonstrate the reward function's effectiveness in achieving both high SE and fair resource distribution. We further enhance reward convergence speed through an improved policy network that boosts user scheduling performance while accelerating reward convergence. The proposed model stabilizes reward curves faster than existing frameworks, enabling quicker convergence on optimal strategies and reducing training time, thereby enhancing the framework's suitability for dynamic MU-MIMO scenarios. Additionally, integrating digital and analog maximal ratio combining (MRC) and zero-forcing (ZF) beamforming techniques offers practical, scalable solutions tailored to future massive MU-MIMO systems.