With the explosive growth of heterogeneous multimodal data across modern industries, efficient compression methods that preserve data integrity are crucial for applications in fields like medical imaging, finance, and remote sensing. Existing compression methods often struggle to handle the varying statistical characteristics of diverse data types, such as text, images, audio, and video. This paper presents MFDNet-ULC, a universal lossless compression framework designed to address these challenges by incorporating both statistical features and semantic content representations. As shown in Figure 1, the proposed framework consists of three key components: a feature extraction module that processes modality-specific data, an attention-guided fusion module that integrates features from different modalities, and a hybrid sequence predictor combining Liquid Neural Networks (LNN) and Transformers to capture both local dynamics and long-range dependencies. The attention-guided fusion module effectively aligns statistical and semantic features, ensuring a robust compression strategy across different data modalities. The hybrid sequence model captures both fine-grained local correlations and global long-range dependencies, improving compression efficiency. Experimental results across multiple datasets demonstrate that MFDNet-ULC outperforms traditional and learning-based methods, achieving up to 75 % improvement in compression ratio and$20 \times$faster compression speeds, making it a robust and efficient solution for real-world compression applications.

The growing complexity of Internet of Things (IoT) systems, especially in Digital Twin (DT) applications, demands data compression methods that ensure real-time transmission and processing. Traditional compression techniques often struggle to meet the heterogeneous, low-latency, and resource-constrained requirements of IoT environments. In this paper, we propose Adaptive Dzip Transformer Compression Technology (ADTCT), a novel framework designed to address these challenges. ADTCT integrates deep neural networks with arithmetic coding, offering significant improvements in compression efficiency while maintaining low latency and minimal memory usage—critical for resource-constrained IoT devices. The core innovation of ADTCT is its dual-stream neural network architecture, which combines a Bootstrap Neural Network (BNN) and a Supporter Neural Network (SNN), coupled with a hybrid training strategy that allows real-time adaptability without relying on external training data. Experimental results demonstrate that ADTCT achieves a compression ratio improvement of 30% over traditional methods such as Gzip, with processing times that are comparable to existing algorithms. These results establish ADTCT as a scalable, reliable solution for real-time data compression across diverse multimodal datasets, including text, images, and sensor data, making it suitable for a variety of IoT applications.

Satellite edge computing can provide ubiquitous and reliable connectivity to remote or disaster area networks that are difficult to serve. However, due to the explosive growth of Internet-of-Things (IoT) data traffic, satellite edge services alone make it difficult to meet the latency and energy demands of abundant IoT devices. Edge learning, combined with edge computing and machine learning, is expected to be the key to solving this problem. This paper constructs a Satellite-assisted IoT network model consisting of terminals, satellites, and a cloud center. Terminals can offload tasks according to actual needs for load balancing. An edge learning approach using cloud edge collaboration is proposed. Then, for concurrent random tasks with different service demands, a minimization problem for the weighted sum of system latency and energy consumption is formulated. Finally, a Model-assisted Two-tier Reinforcement Learning (MTRL) optimization algorithm for task offloading decisions and resource allocation is proposed to solve this problem. Simulation results show that the proposed algorithm has a good performance in convergence. The system performance is better than existing algorithms, which can lead to a 27.5% reduction in the system cost. Furthermore, the proposed algorithm can lead to a smaller increase in system cost as the number of IoT devices increases.

Arbitrary-scale image super-resolution (SR) is often tackled using the implicit neural representation (INR) approach, which relies on a position encoding scheme to im-prove its representation ability. In this paper, we introduce orthogonal position encoding (OPE), an extension of po-sition encoding, and an OPE-Upscale module to replace the INR-based upsampling module for arbitrary-scale im-age super-resolution. Our OPE-Upscale module takes 2D coordinates and latent code as inputs, just like INR, but does not require any training parameters. This parameter-free feature allows the OPE-Upscale module to directly perform linear combination operations, resulting in con-tinuous image reconstruction and achieving arbitrary-scale image reconstruction. As a concise SR framework, our method is computationally efficient and consumes less mem-ory than state-of-the-art methods, as confirmed by exten-sive experiments and evaluations. In addition, our method achieves comparable results with state-of-the-art methods in arbitrary-scale image super-resolution. Lastly, we show that OPE corresponds to a set of orthogonal basis, validating our design principle.11Project page: https://github.com/gaochao-s/ope-sr

In this paper, the problem of optimal passive beamforming design is studied for a reconfigurable intelligent surface (RIS) assisted full-duplex (FD) communication system. In the studied model, two devices communicate with each other using one RIS under the FD mode. Each of the device will receive not only the message from the other device but also the self-interference. The main problem of this work is to minimize the sum transmit power by jointly optimizing the reflection coefficients matrix and the transmit power of devices. To solve this problem, a dual method is proposed, where the dual problem is formulated as a semidefinite programming problem. After solving the dual problem, the phase beamforming of the RIS is obtained in the closed form. Simulation results show that the proposed scheme can reduce up to 66% sum transmit power compared to a conventional RIS assisted half-duplex mode.

This article investigates the problem of resource allocation for multiuser communication networks with a reconfigurable intelligent surface (RIS)-assisted wireless transmitter. In this network, the sum transmit power of the network is minimized by controlling the phase beamforming of the RIS and transmit power of the base station. This problem is posed as a joint optimization problem of transmit power and RIS control, whose goal is to minimize the sum transmit power under signal-to-interference-plus-noise ratio (SINR) constraints of the users. To solve this problem, a dual method is proposed, where the dual problem is obtained as a semidefinite programming problem. After solving the dual problem, the phase beamforming of the RIS is obtained in the closed form, while the optimal transmit power is obtained by using the standard interference function. Simulation results show that the proposed scheme can reduce up to 94% and 27% sum transmit power compared to the maximum ratio transmission (MRT) beamforming and zero-forcing (ZF) beamforming techniques, respectively.

With the rapid development of cloud computing, the user-centric networks with the baseband unit pool have attracted a great deal of attentions in academic and industrial fields. However, limited fronthaul capacity and computation resource have become the bottlenecks inevitably. Thus, this paper investigates the power-efficient transmission in user-centric networks by considering both fronthaul capacity and computation resource constraints, where multiple access points (APs) and user equipments (UEs) are distributed. Specifically, a joint optimization of the beamforming vectors, AP-UE association strategy and transmission time is proposed to minimize the total power consumption (TPC). The formulated mixed integer non-linear problem (MINLP) is NP-hard. To address this problem, the MINLP is first transformed into a convex one via the successive convex approximation and semidefinite relaxation methods. Then, an iterative but effective algorithm is designed by using the property of the solution and applying the Lagrangian dual method. Simulation results show that the proposed algorithm converges rapidly and outperforms benchmark algorithms in terms of TPC.

This paper investigates the dynamic access point (AP) clustering and precoding problem in the downlink of user-centric virtual cell networks. The goal is to maximize the weighted sum spectral efficiency (SE) while satisfying the power constraints and AP clustering constraints in adjacent time slots (TSs). By adopting the random walk mobility to model the mobile user equipments' movement behaviors, we consider dynamic and time-varying channel conditions. Therefore, the weighted sum SE maximization programming takes the form of discrete-time sequence of mixed-integer non-convex optimization problems. In this paper, we propose to solve this sequential problem in two stages. In the first stage, a dynamic AP clustering approach based on discrete particle swarm optimization is developed. This approach takes the advantage of the channel correlation by exploiting the relationship between AP clustering solutions in adjacent TSs to improve the SE performance and reduce complexity. In the second stage, given the AP clustering solution obtained in the first stage, a distributed precoding algorithm is devised via applying the weighted minimum mean square error method. By combining these two stages, we propose a novel dynamic AP clustering and precoding algorithm (DAPC-Pre). The effectiveness of the proposed DAPC-Pre algorithm is verified by the simulation results. In particular, the proposed algorithm converges fast and significantly outperforms benchmark algorithms in terms of sum SE under different dynamic environments.

This paper analyzes the ergodic capacity of a user-centric virtual cell (VC) dense network, where multiple access points (APs) form a VC for each user equipment (UE) and transmit data cooperatively over millimeter wave (mmWave) channels. Different from traditional microwave radio communications, blockage phenomena have an important effect on mmWave transmissions. Accordingly, we adopt a distance-dependent line- of-sight (LOS) probability function and model the locations of the LOS and non-line-of-sight (NLOS) APs as two independent non-homogeneous Poisson point processes (PPP). Invoking this model in a VC dense network, new expressions are derived for the downlink ergodic capacity, accounting for: blockage, small-scale fading and AP cooperation. In particular, we compare the ergodic capacity for different types of fading distributions, including Rayleigh and Nakagami. Numerical results validate our analytical expressions and show that AP cooperation can provide notable capacity gain, especially in low- AP-density regions.

This paper investigates the physical layer security issue of a device-to-device (D2D) underlaid cellular system with a multi-antenna base station (BS) and a multi-antenna eavesdropper. To investigate the potential of D2D communication in improving network security, the conventional network without D2D users (DUs) is first considered. It is shown that the problem of maximizing the sum secrecy rate (SR) of cellular users (CUs) for this special case can be transformed to an assignment problem and optimally solved. Then, a D2D underlaid network is considered. Since the joint optimization of resource block (RB) allocation, CU-DU matching and power control is a mixed integer programming, the problem is difficult to handle. Hence, the RB assignment process is first conducted by ignoring D2D communication, and an iterative algorithm is then proposed to solve the remaining problem. Simulation results show that the sum SR of CUs can be greatly increased by D2D communication, and compared with the existing schemes, a better secrecy performance can be obtained by the proposed algorithms.

In this paper, we consider the network utility maximization problem with various user priorities via jointly optimizing user association, load distribution, and power control in a load-coupled heterogeneous network. In order to tackle the nonconvexity of the problem, we first analyze the problem by obtaining the optimal resource allocation strategy in closed form and characterizing the optimal base station load distribution pattern. Both observations are shown essential in simplifying the original problem and making it possible to transform the nonconvex load distribution and power control problem into convex reformulation via exponential variable transformation. An iterative algorithm with low complexity is accordingly presented to obtain a suboptimal solution to the joint optimization problem. Simulation results show that the proposed algorithm achieves better performance than conventional approaches.

Full‐duplex relay‐assisted device‐to‐device (D2D) communication underlaying fifth generation cellular networks allow devices to exchange information directly and extend coverage via relay strategy. In this study, the authors consider such a scenario, where the D2D communications assisted by fixed‐location full‐duplex relays in interference existing circumstance. Different from previous works, they assume that there are two types of users, cellular users and D2D users. They investigate power control problem and coverage probability performance in the previously assumed situation. Therefore, an effective power control scheme is of great importance to suppress interference between D2D and cellular communications, which can improve the total system throughput and spectral efficiency. To describe it, they formulate a power control optimisation problem for cellular communication and propose a simple on–off power control algorithm for D2D communication. They also obtain an analytic expression for the coverage probability of the cellular link using stochastic geometry according to the proposed algorithm. Simulation results follow to show the rates of both cellular and D2D links in the various numbers of D2D transceivers.

In this paper, we investigate energy-efficient hybrid precoding which consists of baseband precoding and radio frequency (RF) precoding for Millimeter Wave (mmWave) systems over multi-input multi-output (MIMO) interference channels. The considered optimization problem is intractable because both the object function and the constraint are non-convex. Instead of solving the multivariate optimization problem directly, we propose a near-optimal algorithm which includes two steps. First, the original problem is reformulated into an equivalent univariate optimization problem. Second, a near-optimal hybrid precoder is obtained via an orthogonal matching pursuit (OMP) based algorithm. Numerical results verify that the proposed hybrid precoding algorithm achieves a near-optimal EE performance. Moreover, the energy efficiency (EE) maximization algorithm outperforms the sum rate maximization algorithm in terms of the EE performance, especially at high transmit power.

This paper studies pilot reuse and data transmit power control in a D2D underlay massive MIMO system over fading channels. In order to reduce the length of pilots, we propose to reuse a set of orthogonal pilots among DUEs, and the graph coloring based pilot allocation (GCPA) algorithm is utilized to allocate pilots to DUEs. Linear minimum mean square error (LMMSE) filters are used for signal detection. We then derive the lower bound of D2D links' average signal-to-interference-plus-noise ratio (SINR), and formulate a power control problem to minimize D2D links' data transmit power under the target SINR constraints. An iterative method converging to the unique optimal solution is proposed. Simulation results show that the analytical lower bound of the average SINR closely matches the simulated average SINR. What's more, pilot resources can be saved greatly by pilot reuse, and the effect of pilot contamination to the system can be almost neglected by allocating proper number of pilots to DUEs and applying GCPA algorithm.

In this paper, we consider the problem of maximizing the sum energy efficiency (EE) for LTE networks where the interferences occur between cells. Cell load, transmit rate and transmit power, where cell load demonstrates the average resource usage in a cell, are considered in the signal-to- interference-and-noise-ratio (SINR) model. Exploiting the properties of sum EE, we prove that operating at full load is optimal and provide a distributed power control algorithm. With all other powers fixed, we transform the original nonconvex optimization problem in fractional form into an equivalent optimization problem in subtractive form. In high SINR situation, the transformed problem in subtractive form is proved a convex problem. Numerical results demonstrate the remarkable improvements in terms of EE.