25 September 2020 in IEEE Transactions on Wireless Communications
Targeting at a better design of the analogue network coding (ANC)-assisted two-way amplify-and-forward (AF) multiple-input multiple-output (MIMO) multi-relay communication systems, we bring in the nonlinear minimal mean-squared error (MMSE)-decision feedback equalization (DFE) receiving technique to jointly optimize the source precoding, relay amplifying, feed-forward and feedback matrices. Under the transmission power constraints at both source nodes and each relay node, the two-way sum mean-squared error (MSE) of the signal waveform estimation of all data streams is minimized. To solve the complicated nonconvex optimization problem with four groups of system parameters, this paper develops an iterative block coordinate descent (BCD) algorithm, which converges to at least a Nash point. On the basis of it, for mitigating the error propagation in MMSE-DFE receivers, a group of permutation matrix variables, determining the detection orders of all data streams, are further introduced in our system optimization. Moreover, in case there is no sufficiently precise channel state information (CSI), we also make an extension of the developed algorithms, yielding a robust design scheme, to handle the channel uncertainties. Numerical simulation results show that, compared with the existing linear MMSE receiving-based algorithm, our proposed nonlinear ones provide improved MSE and bit-error-rate (BER) performance as well as good robustness against the imperfect CSI, indicating a promising application prospect of this research.
4 October 2021 in IEEE Communications Letters
This letter focuses on the design of multiuser uplink multiple-input multiple-output (MIMO) communications assisted by multiple reconfigurable intelligent surfaces (RISs), where we consider both linear minimal mean-squared error (MMSE) and nonlinear MMSE-decision feedback equalization (DFE) receivers. The RISs, consisting of massive low-cost passive elements, can reflect the incident signals with either continuous or discrete phase shifts. Adopting the sum mean-squared error (MSE) minimization criterion, we jointly optimize the RIS phase shifts, the linear or nonlinear receiving matrices, and the source precoding matrices in iterative manner. Simulation results show that, when the signals in direct user-access point (AP) links suffer severe attenuation, the introduction of indirect user-RIS-AP links can significantly improve the system performance.
25 March 2019 in IEEE Access
To achieve a better long source-destination distance communication in uplink multiaccess scenarios, we propose a multiuser multi-hop amplify-and-forward (AF) multiple-input multiple-output (MIMO) relay technique with nonlinear minimal mean-squared error (MMSE)-decision feedback equalization (DFE) receiver. Under transmission power constraints, this paper focuses on the improvement of reliability, meanwhile, which doesn't lose the effectiveness or require higher complexity. We demonstrate that the optimal structures of relay amplifying matrices lead to a cascading construction for the mean-squared error (MSE) matrices of respective signal waveform estimations at the destination and each relay node. Hence, in (moderately) high signal-to-noise ratio (SNR) environment, the intractable nonconvex optimization problem can be decomposed into easier subproblems for separate optimizations of source precoding and relay amplifying matrices. The source precoding matrix, along with the decision feedback matrix, is obtained by an iterative process, which can converge to a Nash point within the reasonable time. As for the relay amplifying matrices, closed-form water-filling solutions are derived. The simulation and analysis results show that compared to other existing algorithms, which also utilize decomposition methods to simplify operations, the proposed algorithms have better MSE and bit-error-rate (BER) performance without increasing the computing time or signaling overhead, thus providing a new step forward in MIMO relay system design.