Signal processing in multiuser communications
Abstract
This dissertation investigates some signal processing problems for communication systems that can be modeled as linear, multi-input multi-output (MIMO) systems. While such models naturally encompass multi-user communications, they can also represent block encoded communications over a single link. The canonical problem for such MIMO system design is the suppression of interference from other users sharing the same spectrum as in multi-user scenarios (co-channel interference) or from nearby channels due to spectral overlap (adjacent channel interference) for block signalling over a single-input, single-output system. This dissertation makes a contribution to three different interference suppression/cancellation problems. The mean squared error (MSE) is used as a measure of system performance for joint optimization of the transmit-receive filters, subject to a total transmit power constraint. A decomposition of the transmitter and receiver converts the problem to optimization of an equivalent set of decoupled single user systems. An algorithmic solution for the optimized linear system is outlined, and performance evaluation undertaken for some typical scenarios. Next, a similar optimization is attempted for a MIMO system equipped with a decision feedback (DF) receiver. The determinant of the error covariance matrix, termed geometric MSE, is used as the measure of system performance for analytical tractability. Although the mathematical problems for the linear and DF systems may appear to be dissimilar, it is shown that in both cases, the optimized system shares some structural similarities. The final problem considered concerns echo cancellation in a full duplex multi-tone modulation system, which is akin to interference suppression in a MIMO system. Design of a data-driven echo canceller is addressed. A new and efficient echo canceller structure is proposed and analyzed. Algorithm performance in both transient and steady state is analyzed and verified via simulation.
Recommended Citation
Jian Yang,
"Signal processing in multiuser communications"
(January 1, 1993).
Dissertations available from ProQuest.
Paper AAI9331862.
http://repository.upenn.edu/dissertations/AAI9331862
View
More
