1 March 2021 in IEEE Transactions on Wireless Communications
This paper studies fast adaptive beamforming optimization for the signal-to-interference-plus-noise ratio balancing problem in a multiuser multiple-input single-output downlink system. Existing deep learning based approaches to predict beamforming rely on the assumption that the training and testing channels follow the same distribution which may not hold in practice. As a result, a trained model may lead to performance deterioration when the testing network environment changes. To deal with this task mismatch issue, we propose two offline adaptive algorithms based on deep transfer learning and meta-learning, which are able to achieve fast adaptation with the limited new labelled data when the testing wireless environment changes. Furthermore, we propose an online algorithm to enhance the adaptation capability of the offline meta algorithm in realistic non-stationary environments. Simulation results demonstrate that the proposed adaptive algorithms achieve much better performance than the direct deep learning algorithm without adaptation in new environments. The meta-learning algorithm outperforms the deep transfer learning algorithm and achieves near optimal performance. In addition, compared to the offline meta-learning algorithm, the proposed online meta-learning algorithm shows superior adaption performance in changing environments.
1 May 2019 in IEEE Transactions on Communications
A novel cooperative non-orthogonal multiple access (NOMA) strategy is proposed, where a full-duplex cell-center user acts as a friendly relay to help a cell-edge user. An advanced energy harvesting technology based on simultaneous wireless information and power transfer (SWIPT) is used at the cellcenter user to harvest energy used at the cooperative stage. We propose a joint design to optimize the power splitting (PS) ratio and the beamforming vectors. The proposed design aims to maximize energy efficiency (EE) of the system while guaranteeing the minimum required target rate of the cell-edge user and the successful decoding rate at the cell-center user. To make the problem tractable, we use Dinkelbachs method to address the fractional functions and the semidefinite relaxation (SDR) technique to deal with the rank constraint. Then, an iterative algorithm based on successive convex approximation (SCA) is proposed to finally solve the reformulated problem. In order to rich the application of the proposed power allocation algorithm, we extend the proposed strategy to an imperfect channel state information (CSI) mode. S-Procedure is introduced to approximate the channel uncertainties. Numerical results reveal that our proposed iterative algorithms have a good convergence rate. In addition, the proposed strategy offers significant a increase in energy efficiency compared to the existing strategies, especially at the low power region.
17 October 2019 in IEEE Transactions on Vehicular Technology
9 July 2021 in IEEE Transactions on Wireless Communications
This paper studies the fast adaptive beamforming for the multiuser multiple-input single-output downlink. Existing deep learning-based approaches assume that training and testing channels follow the same distribution which causes task mismatch, when the testing environment changes. Although meta learning can deal with the task mismatch, it relies on labelled data and incurs high complexity in the pre-training and fine tuning stages. We propose a simple yet effective adaptive framework to solve the mismatch issue, which trains an embedding model as a transferable feature extractor, followed by fitting the support vector regression. Compared to the existing meta learning algorithm, our method does not necessarily need labelled data in the pre-training and does not need fine-tuning of the pre-trained model in the adaptation. The effectiveness of the proposed method is verified through two well-known applications, i.e., the signal to interference plus noise ratio balancing problem and the sum rate maximization problem. Furthermore, we extend our proposed method to online scenarios in non-stationary environments. Simulation results demonstrate the advantages of the proposed algorithm in terms of both performance and complexity. The proposed framework can also be applied to general radio resource management problems.
20 June 2017 in IEEE Transactions on Vehicular Technology
This paper investigates an outage-constrained secrecy rate maximization (OC-SRM) problem based on the statistical channel uncertainty assumption in an underlay multiple-input multiple-output (MIMO) cognitive radio network, where the secondary transmitter provides simultaneous wireless information and power transfer to receivers. Our objective is to design the transmit covariance matrix and artificial noise aided covariance matrix through maximizing the secrecy rate of the secondary user while satisfying the given outage probability requirements. The designed problem is nonconvex and challenging. We employ the existing theory to reformulate the original problem into two equivalent subproblems by introducing auxiliary variables in order to overcome the difficulty arising from the Shannon capacity expression. The Bernstein-type inequality approach is resorted to conservatively approximate the probabilistic constraints. The merit of the transformation and conversion is that the tractable solutions of the original OC-SRM problem can be easily obtained through solving two convex conic subproblems alternately. Furthermore, we extend the proposed algorithm to solve full uncertainty model design. Numerical results demonstrate the efficacy of the proposed designs.
10 January 2019 in IEEE Transactions on Vehicular Technology
This paper investigates the effect of imperfect channel state information on the performance of the cooperative non-orthogonal multiple access (NOMA) transmission scheme, in which the cell-center user acts as a decode-and-forward relay to assist the information transmission. The cell-center user adopts simultaneous wireless information and power transfer (SWIPT) technique to harvest energy for information forwarding. Based on two channel error models, a maximization problem is formulated to jointly design the robust beamforming and the power splitting ratio while satisfying the minimum data rate for the far user and the successfully decoding requirement for the information of the far user at the near user. For the worst-case design under the deterministic error model, successive convex approximation and the semidefinite relaxation technique are utilized to approximate the non-convex problem to an iterative convex problem. For the outage-constrained design under the stochastic error model, Bernstein-type inequality-based and Large Deviation Inequalitybased approaches are used to safely approximate the probabilistic constraints of the inner-level problem to the deterministic constraints and golden section search algorithm is employed to find out the optimal single variable of the outer-level problem. Furthermore, the rank proof is provided to prove that the relaxation is tight.