Deep learning (DL) is attracting considerable attention in the design of communication systems. This paper derives a deep unfolded conjugate gradient (CG) architecture for large-scale multiple-inputmultiple-output de...
详细信息
Deep learning (DL) is attracting considerable attention in the design of communication systems. This paper derives a deep unfolded conjugate gradient (CG) architecture for large-scale multiple-input multiple-output detection. The proposed technique combines the advantages of a model-driven approach in readily incorporating domain knowledge and deep learning in effective parameters learning. The parameters are trained via backpropagation over a data flow graph inspired from the iterative conjugate gradient method. We derive the closed-form expressions for the gradients for parameters training and discuss early results on the performance in a statistically identical and independent distributed channel where the training overhead is considerably low. It is worth noting that the loss function is based on the residual error that is not an explicit function of the desired signal, which makes the proposed algorithm blind. As an initial framework, we will point to the inherent issues and future directions.
In this paper, an efficient multiple-inputmultiple-output (MIMO) detection scheme with low complexity is proposed. The proposed scheme has a feature of combined a QRD-M, which is the M-algorithm combined with QR deco...
详细信息
ISBN:
(纸本)9781424480166
In this paper, an efficient multiple-inputmultiple-output (MIMO) detection scheme with low complexity is proposed. The proposed scheme has a feature of combined a QRD-M, which is the M-algorithm combined with QR decomposition, with a complex lattice reduction (LR)-aided detection scheme. For the first T stages, the QRD-M detection is executed. And then, the complex LR-aided detection is executed for last N-t - T stages. Simulation results show that the proposed detection scheme provides comparable performance to the QRD-M detection. And this scheme can significantly reduce the computational complexity compared with the QRD-M, because the complexity for the QRD-M is limited by the newly adopted parameter T. The value of T is determined by required system performance. Moreover, to detect the symbols accurately which experienced bad channel, reverse ordering is applied to the proposed scheme. In the case of reverse ordering, the proposed scheme can achieve more improved performance.
暂无评论