Thesis
In wireless communications, there has always been a quest for being able to transmit large amounts of information at the expense of minimal resource utilization. The study of capacity/rate both for single-user and multi-user channels without initial assumption of channel state information (CSI) is important as this capacity not only indicates the upper bound of the rate limit achievable by any transmission scheme but also shows power efficient and inefficient regimes of operation.
For the channels without any assumption of CSI, the capacity analysis could become highly cumbersome even for channel models of moderate complexity. Due to intractability of this analysis, we focus on the high signal-to-noise ratio (SNR) regime in the first part of this thesis which lets us characterize the dominant capacity term, the term with log(P) where P is the transmit power. For single-user single-antenna symbol-by-symbol stationary channels, the exact pre-log (the coefficient of log(P)) is specified for underspread channels. A novel transmission scheme is shown to achieve non-zero pre-log for overspread channels. Considering multi-user downlink (DL) channels with a multi-antenna base station (BS), tight lower and upper bounds of the multiplexing gain are derived for relatively simple block-stationary channels with no assumption of CSI. This analysis also reveals how the quality of CSI at the BS should be scaled with SNR to preserve the multiplexing gain.
A multi-user DL channel shows promising gains even with single-antenna user terminals. Achievability of these gains requires the presence of good quality CSI at the BS which might involve a significant overhead on the uplink (UL) resource. In the second part of this thesis, the focus is on the design and optimization of CSI feedback at the transmitter. We study a multi-user DL system without any assumption of CSI and derive sum rate bounds when CSI acquisition is completely accounted for. These bounds allow maximizing the sum rate by achieving the cost-benefit trade-off of CSI feedback. Further, we propose a novel CSI feedback acquisition strategy for reciprocal channels which combines the use of training sequence and quantized feedback contrary to the classical scheme for such channels where only pilot sequences are used. The results show the superiority of the new hybrid acquisition over traditional CSI acquisition schemes.
For multi-user channels with CSI, the sum rate maximization under fixed power constraints has been widely studied. The dual of this problem, namely, the minimization of transmit power required to achieve specific rate/quality targets at the users' side is also an equally important design problem and of high interest to service providers. In the third part of this thesis, we study the problem of transmit power minimization in conjunction with user scheduling for various user selection schemes. Some analytical results are derived in the limiting case and it is shown that semi-orthogonal greedy user selection performs much better than some other user selection schemes.
Type:
Thesis
Date:
2010-01-27
Department:
Communication systems
Eurecom Ref:
2985
Copyright:
© Université de Nice. Personal use of this material is permitted. The definitive version of this paper was published in Thesis and is available at :
See also:
