Weijun Tang

Weijun Tang

Auto-generated profile, based on publicly available information

Weijun Tang

School of Electronic and Information Engineering, South China University of Technology, Guangzhou, China

Research Focus & Metrics

Academic Career Length

11Years (2016-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
14 September 2020 in IEEE Transactions on Wireless Communications
Verified
Journal Article
Closed access
HFHao Fu
Hao Fu
SFSuili Feng
Suili Feng
WTWeijun Tang
Weijun Tang
DNDerrick Wing Kwan Ng
Derrick Wing Kwan Ng
In this paper, we consider max-min fairness for a downlink two-user multi-input single-output (MISO) system with imperfect channel state information available at transmitter (CSIT) taking into account the total power constraint and the physical layer security. Considering the worst-case channel uncertainty for a potential eavesdropper (PE), we study the robust secure beamforming algorithm design which maximizes the minimum achieved worst-case secrecy rate among single-antenna legitimate users. In contrast to existing schemes adopted in the literatures, we propose a rather unorthodox rate splitting (RS) scheme which advocates the dual use of a common message serving both as a desired message and artificial noise (AN) for legitimate users and the PE, respectively. The algorithm design is formulated as a non-convex optimization problem which is generally intractable. As a compromise approach, we apply the successive convex approximation (SCA) method which facilitates the design of a low-complexity suboptimal iterative algorithm. In each iteration, a rank-constrained semi-definite program (SDP) is solved optimally by SDP relaxation (SDR). Simulation results demonstrate that our proposed robust secure beamforming scheme in the MISO-RS secure transmission system outperforms that of the non-robust counterpart. Moreover, our results also unveil that the proposed RS scheme can achieve a superior performance compared to the existing non-orthogonal multiple access (NOMA) schemes and the traditional scheme.

26 September 2017 in IEEE Transactions on Wireless Communications
Verified
Journal Article
Closed access
WTWeijun Tang
Weijun Tang
SFSuili Feng
Suili Feng
YDYuehua Ding
Yuehua Ding
YLYuan Liu
Yuan Liu
In this paper, we enhance physical layer security for downlink heterogeneous networks by using friendly jammers and full-duplex users. The jammers are selected to transmit jamming signal if their interfering power on the scheduled users is below a threshold, meanwhile the scheduled users confound the eavesdroppers using artificial noise by full-duplexing. Using the tools of stochastic geometry, we derive the expressions of connection probability and secrecy probability. In particular, the locations of active jammers are modeled by a modified Poisson hole process. Determining the jammer selection threshold is further investigated for connection probability maximization subject to the security constraints. A greedy algorithm is proposed to efficiently solve this problem. The accuracy of the theoretical analysis and the efficiency of the proposed algorithm are evaluated by numerical simulations.

22 May 2020 in IEEE Communications Letters
Verified
Journal Article
Closed access
YLYuan Liu
Yuan Liu
ZZZhi Zeng
Zhi Zeng
WTWeijun Tang
Weijun Tang
Fangjiong ChenFangjiong Chen
Fangjiong Chen
The rich mobile data and edge computing enabled wireless networks motivate to deploy artificial intelligence (AI) at network edge, known as edge AI, which integrates wireless communication and machine learning. In communication, data bits are equally important, while in machine learning some data bits are more important. Therefore we can allocate more radio resources to the more important data and allocate less radio resources to the less important data, so as to efficiently utilize the limited radio resources. To this end, how to define “more or less important” of data is the key problem. In this article, we propose two importance criteria to differentiate data’s importance based on their effects on machine learning, one for centralized edge machine learning and the other for distributed edge machine learning. Then, the corresponding radio resource allocation schemes are proposed to improve performance of machine learning. Extensive experiments are conducted for verifying the effectiveness of the proposed data-importance aware radio resource allocation schemes.

30 October 2020 in IEEE Access
Verified
Journal Article
Open access
YZYafang Zhang
Yafang Zhang
SFSuili Feng
Suili Feng
WTWeijun Tang
Weijun Tang
Wireless power transfer (WPT) is an effective way to prolong the lifetime of the energy-constraint networks. In this paper, we investigate a wireless powered cooperative non-orthogonal multiple access (WP-CNOMA) system, consisting of a power beacon (PB), an information transmitter (S), multiple relays (R) and two information receiving devices with near device $D_{1}$ and far device $D_{2}$ . We assume both S and R are energy-constraint and there is no direct link between S and $D_{2}$ . With the help of PB, S and R can harvest energy from it to restart the communication for WP-CNOMA network. For such a system, low-complexity but effective relay and antenna selection schemes are applied. To characterize the performance, outage probabilities and average throughput are derived for linear and non-linear energy harvesting (EH) models, respectively. Moreover, to maximize the average throughput, invoking the unimodal feature for average throughput with respect to the EH time, we find the optimal EH time via Golden section search method. Simulation results validate the accuracy of analytical results, and reveal the performance gain for our system over the benchmark schemes. Also, it can be seen that the non-linear EH model shows different outage behaviors from the linear one. On the other hand, considering the practical application and to improve the performance, the optimization for a simple WP-CNOMA system with single-antenna PB and single relay is also investigated, in which we aim to maximize the minimum throughput by jointly optimizing EH time and power allocation. A low-complexity analytical method is developed to find the max-min rate. Numerical results show that through optimization, the system performance can be improved significantly.

18 March 2019 in IEEE Transactions on Signal Processing
Verified
Journal Article
Closed access
WTWeijun Tang
Weijun Tang
SFSuili Feng
Suili Feng
In this paper, we investigate the network power consumption (NPC) minimization in full-duplex cloud radio access networks (C-RANs) with quality of service (QoS) requirements. We jointly optimize the beamforming vectors, the users' transmit power, the fronthaul compression ratio and the set of transmitting remote radio heads (RRHs). Users' requirements of signal to interference and noise ratio (SINR), per-RRH and user power constraints as well as fronthaul capacity constraints are considered. Due to these conflicting constraints, the optimization problem may be infeasible. Thus, we solve this problem in two steps. In Step I, a minimum mean square error (MMSE) based user selection algorithm is proposed to find the largest subset of feasible users. In Step II, an algorithm based on the reweighted $l_1$ -norm minimization method is proposed to solve the NPC problem with the users selected in Step I. The solutions obtained by the proposed algorithms are proved to monotonically converge and satisfy the Karush-Kuhn-Tucker (KKT) conditions. Moreover, our proposed algorithms are applicable to both fullduplex and half-duplex C-RANs. We comprehensively compare these two types of C-RANs and propose some valuable insights for system design.

14 January 2021 in IEEE Wireless Communications Letters
Verified
Journal Article
Closed access
ZZZhi Zeng
Zhi Zeng
YLYuan Liu
Yuan Liu
WTWeijun Tang
Weijun Tang
Fangjiong ChenFangjiong Chen
Fangjiong Chen
Edge intelligence requires to fast access distributed data samples generated by edge devices. The challenge is using limited radio resource to acquire massive data samples for training machine learning models at edge server. In this letter, we propose a new communication-efficient edge intelligence scheme where the most useful data samples are selected to train the model. Here the usefulness or values of data samples is measured by data diversity which is defined as the difference between data samples. We derive a close-form expression of data diversity that combines data informativeness and channel quality. Then a joint data-and-channel diversity aware multiuser scheduling algorithm is proposed. We find that noise is useful for enhancing data diversity under some conditions.