Semantic communication has emerged as a promising paradigm for next-generation wireless networks, offering enhanced efficiency by reducing transmission data. However, implementing semantic communication faces significant challenges, particularly in resource-constrained devices that cannot support the complex artificial intelligence (AI) models required for semantic extraction. This paper addresses this challenge by proposing a novel unmanned aerial vehicle (UAV)-enabled semantic-bit coexisting relay system, where the UAV serves as intermediate nodes to assist transmissions from resource-limited users to the base station. By deploying semantic extraction models at the UAV, the proposed system solves the computational resource limitations for user devices while minimizing transmission latency via data size reduction. In such a system, we formulate a system latency minimization problem that jointly considers semantic compression model selection and bandwidth allocation. To address this complex problem, we develop an effective solution method by decomposing the original problem into a semantic compression model selection based on performance-latency trade-offs and a bandwidth-allocation optimization via convex optimization techniques. Extensive numerical evaluations demonstrate that the proposed framework consistently outperforms conventional schemes across diverse network settings and compression parameters, significantly reducing end-to-end latency while maintaining high-quality semantic communication.

This research examines quantized uplink multi-user MIMO communication systems with low-resolution quantizers at users and base stations (BS). In such a system, we employ the non-orthogonal multiple access (NOMA) technique for communication between users and the BS to enhance communication performance. To maximize the number of users that satisfy the quality of service (QoS) requirement while minimizing the user’s transmit power, we jointly optimize the transmit power and precoding matrices at the users and the digital beamforming matrix at the BS. Owing to the non-convexity of the objective function, we transform the problem into a reinforcement learning-based problem and propose a deep reinforcement learning (DRL) framework named QNOMA-DRLPA to overcome the challenge. Because the nature of the action decided by the DRL algorithm may not satisfy the problem constraints, we propose a postactor process to redesign the actions to meet all the problem constraints. In the simulation, we assess the proposed framework’s performance in training convergence and demonstrate its superior performance under various environmental parameters compared with other benchmark schemes.

Aerial access networks, comprising a hierarchical model of high-altitude platforms (HAPs) and multiple unmanned aerial vehicles (UAVs), are considered a promising technology to enhance the service experience of Internet of Things Devices (IoTD), especially in underserved areas where terrestrial base stations (TBSs) do not exist. In such scenarios, optimally orchestrating the limited computation, communication, and energy resources in both HAPs and UAVs is crucial toward for an efficient aerial networking infrastructure. Thus, in this study, we investigate and formulate the joint IoTDs association, partial offloading, and communication resource allocations (JAPORAs) decisions problem in heterogeneous Aerial Access IoT (AAIoT) networks to maximize service satisfaction for IoTDs, while minimizing their total energy consumption. In particular, the formulated problem is transformed into a multiagent Markov decision process (MAMDP) to deal with its nonconvexity and environmental dynamicity. To solve the problem, we propose a multiagent policy-gradient-based deep actor–critic algorithm, named MADDPG-JAPORA, with centralized training and decentralized execution. Our extensive numerical experiments demonstrated that MADDPG-JAPORA reliably converges and provides superior performance compared with other state-of-the-art schemes.

In recent years, video has become a tremendous growth of media in the content delivery network (CDN) coupled with the enormous increase of users. However, it leads to new challenges including an explosion of data demand on networking resources, backhaul bottleneck, and many congestions between transmissions in the network. To address these network congestions and minimize the content download latency, an LEO Satellite-based overlay system in CDN is proposed. A Low Earth orbit (LEO) satellite network, where the user equipment (UE) is covered by multiple satellites, is an important solution to the wireless communication network in the future. To ensure the quality of the service of the LEO satellite network, the handover problem and caching problem need to be considered. The paper focuses on a wireless network consisting of various caching nodes to serve users' requests supported by an LEO satellite-based overlay system. Our goal is the maximization the number of UEs covered by a Satellite by considering the location of users and the Received Signal Strength Indicator (RSSI) of users to select the best action. The Deep Reinforcement Learning (DRL) with multi-agent Q learning algorithm and Caching update algorithm with the popularity parameters are proposed to solve Problem statements of our System Model.

Recent studies have shown that rate splitting multiple access (RSMA), which depends on multi-antenna rate splitting (RS) at the transmitter and successive interference cancellation (SIC) at the receivers, successfully controls interference in multi-antenna communication networks. This paper examines RSMA's applicability to video streaming applications in cloud radio access networks (C-RAN). We aim to address a practical challenge to maximize the perceived quality of end users while keeping the delay constraints remained satisfied using RSMA. We propose a learning-based framework to select appropriate video quality together with beamforming vectors according to current defined system state. The simulation figure confirms that the learning behavior of proposed learning scheme is stable.