This paper proposes an energy efficient resource allocation design algorithm for an intelligent reflecting surface (IRS)-assisted downlink ultra-reliable low-latency communication (URLLC) network. This setup features a multi-antenna base station (BS) transmitting data traffic to a group of URLLC users with short packet lengths. We maximize the total network's energy efficiency (EE) through the optimization of active beamformers at the BS and passive beamformers (a.k.a. phase shifts) at the IRS. The main non-convex problem is divided into two sub-problems. An alternating optimization (AO) approach is then used to solve the problem. Through the use of the successive convex approximation (SCA) with a novel iterative rank relaxation method, we construct a concave-convex objective function for each sub-problem. The first sub-problem is a fractional program that is solved using the Dinkelbach method and a penalty-based approach. The second sub-problem is then solved based on semi-definite programming (SDP) and the penalty-based approach. The iterative solution gradually approaches the rank-one for both the active beamforming and unit modulus IRS phase-shift sub-problems. Our results demonstrate the efficacy of the proposed solution compared to existing benchmarks.
This paper considers the energy efficiency (EE) maximization of a simultaneous wireless information and power transfer (SWIPT)-assisted unmanned aerial vehicles (UAV) cooperative network operating at TeraHertz (THz) frequencies. The source performs SWIPT enabling the UAV to receive both power and information while also transmitting the information to a designated destination node. Subsequently, the UAV utilizes the harvested energy to relay the data to the intended destination node effectively. Specifically, we maximize EE by optimizing the non-orthogonal multiple access (NOMA) power allocation coefficients, SWIPT power splitting (PS) ratio, and UAV trajectory. The main problem is broken down into a two-stage optimization problem and solved using an alternating optimization approach. In the first stage, optimization of the PS ratio and trajectory is performed by employing successive convex approximation using a lower bound on the exponential factor in the THz channel model. In the second phase, the NOMA power coefficients are optimized using a quadratic transform approach. Numerical results demonstrate the effectiveness of our proposed resource allocation algorithm compared to the baselines where there is no trajectory optimization or no NOMA power or PS optimization.
In this paper, we first develop a tractable mathematical model of the handoff (HO)-aware rate outage experienced by a typical connected and autonomous vehicle (CAV) in a given THz vehicular network. The derived model captures the impact of line-of-sight (LOS) Nakagami-m fading channels, interference, and molecular absorption effects. We first derive the statistics of the interference-plus-molecular absorption noise ratio and demonstrate that it can be approximated by Gamma distribution using Welch-Satterthwaite approximation. Then, we show that the distribution of signal-to-interference-plus-molecular absorption noise ratio (SINR) follows a generalized Beta prime distribution. Based on this, a closed-form HO-aware rate outage expression is derived. Finally, we formulate and solve a CAVs' traffic flow maximization problem to optimize the base-stations (BSs) density and speed of CAVs with collision avoidance, rate outage, and CAVs' minimum traffic flow constraint. The CAVs' traffic flow is modeled using Log-Normal distribution. Our numerical results validate the accuracy of the derived expressions using Monte-Carlo simulations and discuss useful insights related to optimal BS density and CAVs' speed as a function of crash intensity level, THz molecular absorption effects, minimum road-traffic flow and rate requirements, and maximum speed and rate outage limits.
This paper presents a deep reinforcement learning solution for optimizing multi-UAV cell-association decisions and their moving velocity on a 3D aerial highway. The objective is to enhance transportation and communication performance, including collision avoidance, connectivity, and handovers. The problem is formulated as a Markov decision process (MDP) with UAVs' states defined by velocities and communication data rates. We propose a neural architecture with a shared decision module and multiple network branches, each dedicated to a specific action dimension in a 2D transportation-communication space. This design efficiently handles the multi-dimensional action space, allowing independence for individual action dimensions. We introduce two models, Branching Dueling Q-Network (BDQ) and Branching Dueling Double Deep Q-Network (Dueling DDQN), to demonstrate the approach. Simulation results show a significant improvement of 18.32% compared to existing benchmarks.
Rate splitting multiple access (RSMA) and non-orthogonal multiple access (NOMA) are the key enabling multiple access techniques to enable massive connectivity. However, it is unclear whether RSMA would consistently outperform NOMA from a system sum-rate perspective, users' fairness, as well as convergence and feasibility of the resource allocation solutions. This paper investigates the weighted sum-rate maximization problem to optimize power and rate allocations in a hybrid RSMA-NOMA network. In the hybrid RSMA-NOMA, by optimally allocating the maximum power budget to each scheme, the BS operates on NOMA and RSMA in two orthogonal channels, allowing users to simultaneously receive signals on both RSMA and NOMA. Based on the successive convex approximation (SCA) approach, we jointly optimize the power allocation of users in NOMA and RSMA, the rate allocation of users in RSMA, and the power budget allocation for NOMA and RSMA considering successive interference cancellation (SIC) constraints. Numerical results demonstrate the trade-offs that hybrid RSMA-NOMA access offers in terms of system sum rate, fairness, convergence, and feasibility of the solutions.
Deep neural networks (DNNs) are emerging as a potential solution to solve NP-hard wireless resource allocation problems. However, in the presence of intricate constraints, e.g., users' quality-of-service (QoS) constraints, guaranteeing constraint satisfaction becomes a fundamental challenge. In this paper, we propose a novel unsupervised learning framework to solve the classical power control problem in a multi-user interference channel, where the objective is to maximize the network sumrate under users' minimum data rate or QoS requirements and power budget constraints. Utilizing a differentiable projection function, two novel deep learning (DL) solutions are pursued. The first is called Deep Implicit Projection Network (DIPNet), and the second is called Deep Explicit Projection Network (DEPNet). DIPNet utilizes a differentiable convex optimization layer to implicitly define a projection function. On the other hand, DEPNet uses an explicitly-defined projection function, which has an iterative nature and relies on a differentiable correction process. DIPNet requires convex constraints; whereas, the DEPNet does not require convexity and has a reduced computational complexity. To enhance the sum-rate performance of the proposed models even further, Frank-Wolfe algorithm (FW) has been applied to the output of the proposed models. Extensive simulations depict that the proposed DNN solutions not only improve the achievable data rate but also achieve zero constraint violation probability, compared to the existing DNNs. The proposed solutions outperform the classic optimization methods in terms of computation time complexity.
Over-the-air federated learning (OTA-FL) is an emerging technique to reduce the computation and communication overload at the PS caused by the orthogonal transmissions of the model updates in conventional federated learning (FL). This reduction is achieved at the expense of introducing aggregation error that can be efficiently suppressed by means of receive beamforming via large array-antennas. This paper studies OTA-FL in massive multiple-input multiple-output (MIMO) systems by considering a realistic scenario in which the edge server, despite its large antenna array, is restricted in the number of radio frequency (RF)-chains. For this setting, the beamforming for over-the-air model aggregation needs to be addressed jointly with antenna selection. This leads to an NP-hard problem due to the combinatorial nature of the optimization. We tackle this problem via two different approaches. In the first approach, we use the penalty dual decomposition (PDD) technique to develop a two-tier algorithm for joint antenna selection and beamforming. The second approach interprets the antenna selection task as a sparse recovery problem and develops two iterative joint algorithms based on the Lasso and fast iterative soft-thresholding methods. Convergence and complexity analysis is presented for all the schemes. The numerical investigations depict that the algorithms based on the sparse recovery techniques outperform the PDD-based algorithm, when the number of RF-chains at the edge server is much smaller than its array size. However, as the number of RF-chains increases, the PDD approach starts to be superior. Our simulations further depict that learning performance with all the antennas being active at the PS can be closely tracked by selecting less than 20% of the antennas at the PS.
This paper investigates the problem of minimizing the age-of-information (AoI) and transmit power consumption in a vehicular network, where a roadside unit (RSU) provides timely updates about a set of physical processes to vehicles. Each vehicle is interested in maintaining the freshness of its information status about one or more physical processes. A framework is proposed to optimize the decisions to unicast, multicast, broadcast, or not transmit updates to vehicles as well as power allocations to minimize the AoI and the RSU's power consumption over a time horizon. The formulated problem is a mixed-integer nonlinear programming problem (MINLP), thus a global optimal solution is difficult to achieve. In this context, we first develop an ant colony optimization (ACO) solution which provides near-optimal performance and thus serves as an efficient benchmark. Then, for real-time implementation, we develop a deep reinforcement learning (DRL) framework that captures the vehicles' demands and channel conditions in the state space and assigns processes to vehicles through dynamic unicast-multicast scheduling actions. Complexity analysis of the proposed algorithms is presented. Simulation results depict interesting trade-offs between AoI and power consumption as a function of the network parameters.
Passive beamforming in reconfigurable intelligent surfaces (RISs) enables a feasible and efficient way of communication when the RIS reflection coefficients are precisely adjusted. In this paper, we present a framework to track the RIS reflection coefficients with the aid of deep learning from a time-series prediction perspective in a terahertz (THz) communication system. The proposed framework achieves a two-step enhancement over the similar learning-driven counterparts. Specifically, in the first step, we train a liquid state machine (LSM) to track the historical RIS reflection coefficients at prior time steps (known as a time-series sequence) and predict their upcoming time steps. We also fine-tune the trained LSM through Xavier initialization technique to decrease the prediction variance, thus resulting in a higher prediction accuracy. In the second step, we use ensemble learning technique which leverages on the prediction power of multiple LSMs to minimize the prediction variance and improve the precision of the first step. It is numerically demonstrated that, in the first step, employing the Xavier initialization technique to fine-tune the LSM results in at most 26% lower LSM prediction variance and as much as 46% achievable spectral efficiency (SE) improvement over the existing counterparts, when an RIS of size 11x11 is deployed. In the second step, under the same computational complexity of training a single LSM, the ensemble learning with multiple LSMs degrades the prediction variance of a single LSM up to 66% and improves the system achievable SE at most 54%.