Capacity bounds via duality with applications to multiple-antenna systems on flat-fading channels

Lapidoth, Amos; Moser, Stefan M.

doi:10.1109/tit.2003.817449

Cited by 397 publications

(536 citation statements)

References 23 publications

Supporting

Mentioning

528

Contrasting

Unclassified

Order By: Relevance

“…By Lemma 8 with ψ = κ and A =Ũ, there exists a set U ⊆Ũ such that κ U 5 The function g is measurable by the following argument: κ −1 i is continuous by the inverse function theorem. Hence, for all u with equal F (u), the denominator in the definition of g is just the maximum over a finite set of continuous functions and thus measureable.…”

Section: Lemmamentioning

confidence: 99%

A lower bound on the noncoherent capacity pre-log for the MIMO channel with temporally correlated fading

Koliander

Riegler

Durisi

et al. 2012

2012 50th Annual Allerton Conference on Communication, Control, and Computing (Allerton)

View full text Add to dashboard Cite

Abstract-We derive a lower bound on the capacity pre-log of a temporally correlated Rayleigh block-fading multiple-input multiple-output (MIMO) channel with T transmit antennas and R receive antennas in the noncoherent setting (no a priori channel knowledge at the transmitter and the receiver). In this model, the fading process changes independently across blocks of length L and is temporally correlated within each block for each transmit-receive antenna pair, with a given rank Q of the corresponding correlation matrix. Our result implies that for almost all choices of the coloring matrix that models the temporal correlation, the pre-log can be lower-bounded bythat R is sufficiently large. The widely used constant block-fading model is equivalent to the temporally correlated block-fading model with Q = 1 for the special case when the temporal correlation for each transmit-receive antenna pair is the same, which is unlikely to be observed in practice. For the constant block-fading model, the capacity pre-log is given by T (1 − T /L), which is smaller than our lower bound for the case Q = 1. Thus, our result suggests that the assumptions underlying the constant blockfading model lead to a pessimistic result for the capacity pre-log.

show abstract

Section: Lemmamentioning

confidence: 99%

A lower bound on the noncoherent capacity pre-log for the MIMO channel with temporally correlated fading

Koliander

Riegler

Durisi

et al. 2012

2012 50th Annual Allerton Conference on Communication, Control, and Computing (Allerton)

View full text Add to dashboard Cite

show abstract

“…5 In fact, decibel units are used to quantify the power offset in some numerical examples, but all of the analytical expressions in the paper are given in 3-dB units. 5 The zero-order counterpart of L for channels unknown to the receiver is the fading number introduced in [7]. In contrast to the coherent case that occupies us here, though, in the noncoherent case the leading term in the highexpansion is not necessarily linear in j and, as a result, the interpretation of the fading number does not mirror that of the power offset.…”

Section: High-snr Measuresmentioning

confidence: 98%

“…Rayleigh-faded channel and 1 Coherent operation may be unfeasible for ! 1 when the noise becomes commensurate with the uncertainty in the knowledge of the precise state of the channel and the logarithmic scaling in (1) breaks down [6], [7]. For levels of practical interest, however, the premise of coherent reception is valid in a vast majority of channels and an asymptotic characterization accurately captures the behavior of the capacity at those levels.…”

Section: A Motivationmentioning

confidence: 99%

“…Denoting by and the diagonal eigenvalue matrices of and , 7 we define corresponding and diagonal matrices and obtained by striking out the empty rows and columns of and , respectively. It is also convenient to introduce an matrix-valued function , , whose th entry is (25) where (26) while is the th diagonal entry of the matrix argument .…”

Section: B the High-snr Power Offsetmentioning

confidence: 99%

“…For an isotropic input, is given by the right-hand side of (28) with replaced by the identity matrix. Proof: See Appendix E. 7 Since the eigenspaces of 2 2 2 and 8 8 8 are common, the diagonal entries of 3 3 3…”

Section: Propositionmentioning

confidence: 99%

See 2 more Smart Citations

High-SNR power offset in multi-antenna communication

Lozano¹,

Tulino

Verdú

International Symposium onInformation Theory, 2004. ISIT 2004. Proceedings.

View full text Add to dashboard Cite

Abstract-The analysis of the multiple-antenna capacity in the highregime has hitherto focused on the highslope (or maximum multiplexing gain), which quantifies the multiplicative increase as a function of the number of antennas. This traditional characterization is unable to assess the impact of prominent channel features since, for a majority of channels, the slope equals the minimum of the number of transmit and receive antennas. Furthermore, a characterization based solely on the slope captures only the scaling but it has no notion of the power required for a certain capacity. This paper advocates a more refined characterization whereby, as a function of dB , the high-capacity is expanded as an affine function where the impact of channel features such as antenna correlation, unfaded components, etc., resides in the zero-order term or power offset. The power offset, for which we find insightful closed-form expressions, is shown to play a chief role for levels of practical interest.

show abstract

Iterative channel estimation and decoding in turbo coded space‐time systems

Ronen

Bross

Shamai

et al. 2007

Trans Emerging Tel Tech

View full text Add to dashboard Cite

A noncoherent multiple antenna coded communication system over both block fading and time‐selective flat fading Rayleigh channel is considered. Several iterative decoding strategies consisting of pilot symbol assisted demodulation and turbo decoding are proposed. It is shown that, at a system rate of 0.9 bits per channel use, the SNR required to achieve a BER of 10−4 is only about 2dB from the imperfect CSI capacity lower bound. The channel estimation error variance is demonstrated to be close to the theoretical limit set by the Modified Cramér‐Rao bound. The simulation results coincide with the trends dictated by recent information‐theoretic analysis and demonstrate the effect that inaccuracy of the channel estimation has on the performance of the iterative decoding scheme. Copyright © 2007 John Wiley & Sons, Ltd.

show abstract

Capacity bounds via duality with applications to multiple-antenna systems on flat-fading channels

Cited by 397 publications

References 23 publications

A lower bound on the noncoherent capacity pre-log for the MIMO channel with temporally correlated fading

A lower bound on the noncoherent capacity pre-log for the MIMO channel with temporally correlated fading

High-SNR power offset in multi-antenna communication

Iterative channel estimation and decoding in turbo coded space‐time systems

Contact Info

Product

Resources

About