A Generalized Matched Filter Framework for Cellular Massive MIMO Networks
Final Report Abstract
In this project, we investigated massive MIMO systems, i.e., wireless communication systems where the base station is equipped with a large number of antenna elements, e.g., 100 or even more antennas, developed solutions for both, the uplink and the downlink transmissions, and analyzed the bounds for the possible spectral e ciencies of such systems. The key ingredient of our studies was the employment of the concept of the generalized matched filter which is based on a linear transformation of the channel estimate where the linear transformation is a function of the channel covariance matrices. The main challenge for massive MIMO systems is called pilot contamination where the channel estimates contain systematic errors due to the limited number of available training sequences. In the studies of the project, we observed that the pilot contamination can successfully be suppressed or even eliminated. This observation was suprising since the capabilities of the generalized matched filter merely come from the utilization of the channel covariance matrices. Additionally, this observation was supported by our theoretical study of the high signal-tonoise ratio regime. Since the dimensionality of the problems is very large due to the large number of antenna elements, we also investigated the application of existing approximations exploiting properties due to the antenna array structure to reduce computational complexity. The result of the corresponding studies was twofold. Firstly, the resulting complexity is tremendously reduced. Secondly, the approximation e↵ects the perfomance only marginally. As the generalized matched filter strongly depends on the knowledge of the channel covariance matrices, we also considered the problem of estimating these covariance matrices although the channel estimates contain errors due to pilot contamination. Again, we proposed low complexity solutions for this problem and discussed the allocation of pilot sequences to optimize a network-wide utility function. In the second half of the project, we put our focus on systems with frequency duplex that have a considerable practical importance but lead to the problem that the channel estimates in the uplink cannot be directly used as estimates in the downlink. We investigated a method to transfer the estimates of the channel covariance matrices obtained in the uplink to estimates in the downlink. Moreover, we proposed a method to optimize the whole massive MIMO system not only for an isolated cell taking into account the power restrictions of all base stations.
Publications
-
Joint Covariance Matrix Estimation and Pilot Allocation in Massive MIMO Systems. In 2017 IEEE International Conference on Communications (ICC), 2017
D. Neumann, K. Shibli, M. Joham, and W. Utschick
-
On MSE Based Receiver Design for Massive MIMO. In 11th International ITG Conference on Systems, Communications and Coding (SCC 2017), 2017
D. Neumann, M. Joham, and W. Utschick
-
A Bilinear Equalizer for Massive MIMO Systems. IEEE Transactions on Signal Processing, 66(14):3740–3751, 2018
D. Neumann, T. Wiese, M. Joham, and W. Utschick
-
Channel Covariance Identification in FDD Massive MIMO Systems. In Proc. 2018 IEEE Global Conference on Signal and Information Processing (GlobalSIP), 2018
J. P. González-Coma, P. Suárez-Casal, P. M. Castro, L. Castedo, and M. Joham
-
Covariance Matrix Estimation in Massive MIMO. IEEE Signal Processing Letters, 25(6):863–867, 2018
D. Neumann, M. Joham, and W. Utschick
-
Bilinear Precoding for FDD Massive MIMO Systems with Imperfect Covariance Matrices. In 24th International ITG Workshop on Smart Antennas (WSA 2020), 2020
D. Ben Amor, M. Joham, and W. Utschick
-
Uplink Downlink Duality for Multi-Cell Massive MISO FDD Systems with per Base Station Power Constraints. In 25th International ITG Workshop on Smart Antennas (WSA 2021), 2021
D. Ben Amor, F. Strasser, M. Joham, and W. Utschick
