Hybrid Nonlinear Transceiver Optimization for the RIS-Aided MIMO Downlink

Abstract

The hybrid nonlinear transceiver optimization problem of reconfigurable intelligent surface (RIS)-aided multi-user multiple-input multiple-output (MU-MIMO) downlink is investigated. Specifically, the Tomlinson-Harashima precoder (THP) and the hybrid transmit precoder (TPC) of the base station are jointly optimized with the linear digital receivers of mobile users. The triangular feedback matrix of the THP is optimized and the optimal solution is derived in closed form based on a matrix inequality. Moreover, in order to tackle the nonconvexity of the constant-modulus constraints imposed on the analog TPC, the Majorization-Minimization (MM) based reconfigurable optimization framework is proposed, which strikes a trade-off between the implementation complexity and system performance in a reconfigurable manner. Explicitly, our MM-based reconfigurable optimization framework is capable of optimizing the analog TPC in a dynamically reconfigurable manner on an element-byelement, column-by-column, row-by-row or block-by-block basis. Moreover, an MM-based reconfigurable algorithm is proposed for the optimization of the phase shifting matrix at RIS, which also suffers from constant-modulus constraints. In the proposed MM-based reconfigurable algorithm, the RIS can be partitioned into a series of subarrays for striking different performance vs. complexity tradeoffs. Finally, our numerical results demonstrate the performance advantages of the proposed nonlinear hybrid transceiver optimization techniques.

Low-Complexity Tomlinson-Harashima Precoding Update Algorithm for Massive MIMO System

“…By comparing (20) with (21), we can see that Q t+1 ∈ C M ×(N −1) is a sub-matrix of Q by deleting the last column, and R t+1 ∈ C (N −1)×(N −1) is a sub-matrix of R by deleting the last column and the last row. Then we can easily obtain F t+1 , G t+1 and B t+1 using (3)- (5).…”

“…By minimizing the MMSE of the system under channel uncertainties with realistic transmit power and unit modulus constraints, the proposed design achieved lower complexity and power cost. In [21], a novel algorithm was proposed to jointly optimize the THP and the hybrid transmit precoder of the BS with the linear digital receivers of mobile UEs.…”

Joint Transceiver and Reconfigurable Intelligent Surface Design for Multiuser mmWave MIMO Systems Relying on Non-Diagonal Phase Shift Matrices

“…A key challenge in RIS-aided systems is to jointly design the active beamformer at the BS and the passive beamformer at the RIS. The authors of [7], [8], [9], [10], [11], [12], [13], [14], [15] consider different system models and present algorithms for the joint design of these beamformers. In [7], the authors consider a setup with an access point (AP) that has multiple antennas communicating with multiple user equipment (UEs), each having a single antenna, with the help of an RIS.…”