Font Size: a A A

Distributed Multiuser Detection For Massive MIMO System

Posted on:2016-12-23Degree:DoctorType:Dissertation
Country:ChinaCandidate:Z Q YueFull Text:PDF
GTID:1108330503469636Subject:Information and Communication Engineering
Abstract/Summary:PDF Full Text Request
In 2010, Thomas L. Marzetta proposed the concept of large-scale MIMO(Massive MIMO). Massive MIMO is thought of as antenna arrays with a few hundred antennas at the base station simultaneously serving tens of users in the same time-frequency bins. However, for unfavorable propagation case(Both the number of antennas at the BS and the number of users grow large while their ratio is bounded), conventional MIMO receiver could not meet the requirements of architecture, the performance and complexity simultaneously. Optimal maximum a posteriori(MAP) detector has high computational complexity that grows exponentially with the number of users. Even the suboptimal MMSE-SIC(Minimum Mean Square Error-Successive Interference Cancellation) algorithm requires complexity proportional to the cube of the number of antenna units. Even worse, all data must be collected at a central node before detection. This motivates research for distributed detection schemes, done at each antenna unit separately, with high performance and low complexity.Firstly, this paper proposed a distributed detection scheme done at each antenna unit separately, termed complex Gaussian belief propagation algorithm(CGa BP), for multicell multiuser detection. The multiuser detection problem is reduced to a sequence of scalar AWGN(additive white Gaussian noise) estimations, and detecting each individual user using CGa BP is equivalent to detecting the same user through a scalar Gaussian channel, where the additive noise is due to the collective impact of interfering users. Numerical results show that CGa BP is better than MMSE-SIC for BPSK symbols.Secondly, for the real Gaussian random channel matrix, this paper proposed a method to analyze the performance of CGa BP algorithm in large system limit by the state evolution theory, and shows that the dynamical behavior of CGa BP algorithm can be predicated exactly. Mean square error of individual user estimation using CGa BP achieves the MMSE of the same user in AWGN channel, where the noise is determined by a recursive formula. Empirical validations show that, even for medium problem size, the dynamical behavior of CGa BP algorithm can be predicated regardless of Gaussian or digital modulation signal. Compared with MMSE-SIC receiver, the CGa BP algorithm has lower computational complexity and message passing overhead.Finally, this paper extended the CGa BP algorithm for two different scenarios. First, assuming that the transmit power gain of user is unknown, this paper proposed the adaptive-CGa BP algorithm, which combines the original CGa BP algorithm with the maximum likelihood parameter learning, to estimates transmit power gain blindly while multiuser detection. Simulation results show that at the high SNR or if the number of base station antenna is large, the performance of adaptive-CGa BP algorithm is close to the performance of the original one. And then, assuming that the channel state information is unknown, this paper proposed the CGa BP algorithm with imperfect channel state information. The simulation results show that, although the CGa BP algorithm degrades in the presence of imperfect channel state information and pilot contamination, the algorithm still outperforms the multicell MMSE-SIC algorithm.
Keywords/Search Tags:massive MIMO, multiuser detection, distributed, state evolution, channel state information, pilot contamination
PDF Full Text Request
Related items