Robust Power Allocation in MIMO-NOMA Systems

Generalized singular-value decomposition (GSVD) has been recognized as an enabler of multiple-input–multiple-output (MIMO) nonorthogonal multiple-access (NOMA) systems. The GSVD is leveraged to MIMO-NOMA with channel estimation errors. Channel uncertainty is uniquely incorporated in the variances of elements of error matrices in the signal-to-interference-plus-noise ratio (SINR) balancing problem, and tractable convex formulations are derived for different uncertainty models in this letter. The proposed design is numerically evaluated and its superiority over non-robust and orthogonal multiple access (OMA) systems is established.