Optimized Precoding for MU-MIMO With Fronthaul Quantization

Quantization-Aware Precoding For Mu-Mimo With Limited-Capacity Fronthaul

“…We notice that, despite the approximations made to lower the complexity in the SD-based method, the difference in the sum rate is negligible after convergence. The difference between the sum rate at the starting point and the stationary point approximately shows the improvement of the proposed sum-rate maximization algorithm compared to our previous results in [10], which considered sum MSE minimization. Fig.…”

“…The effect of limited fronthaul capacity is studied in [9], but the precoding design was not quantized. In [10], the authors proposed a fronthaul quantization-aware precoding design that minimizes the sum MSE, which will generally not maximize the sum rate. Many previous studies suggest designing a precoding matrix by maximizing the sum rate, often using the WMMSE approach; see [11,12] and references therein.…”

“…It can be solved using general-purpose methods, such as CVX [14]. By iterating between updating β k using (8), d k using (10), and P by solving P 3 , we obtain a block coordinate descent algorithm that will converge to a stationary point (for the same reasons as in [4]). This algorithm is summarized in Figure 1.…”

“…. , K. Thus, in addition to the more efficient search strategy, the reformulation of problem P 5 also significantly reduces the dimension of each subproblem [10]. By defining V = H H D H DH + λI M , we can obtain the equivalent formulation of each term of the objective function in (16) as…”

“…After we quantize the precoding matrix W ∈ C M ×K to obtain P ∈ P M ×K , we refine the elements sequentially. We consider the second closest quantization levels in both the real and imaginary dimensions according to Euclidean distance [10]. This search gives us three alternative ways of quantizing each element in the precoding matrix W .…”