Beruflich Dokumente
Kultur Dokumente
on Beam-Nulling
Mabruk Gheryani, Zhiyuan Wu, and Yousef R. Shayan
Concordia University, Department of Electrical Engineering
Montreal, Quebec, H4G 2W1, Canada
email: (m gherya, zy wu, yshayan)@ece.concordia.ca
Abstract— In this paper, we propose a scheme called “beam- data rate, such as constellation and coding rate. There are also
nulling” using the same feedback bandwidth as beamforming but other beamforming schemes based on various criteria, such as
with higher capacity. In the beam-nulling scheme, the eigenvector [7]- [14]. Note that the conventional beamforming is optimal
of the weakest subchannel is fed back and then signals are sent
over a generated subspace orthogonal to the weakest subchannel. in terms of maximizing the SNR at the receiver. However,
Hence, the scheme can achieve high capacity. The capacities of it is sub-optimal from a MIMO capacity point of view, since
water-filling, equal power, beamforming and beam-nulling are only one data stream, instead of parallel streams, is transmitted
compared through theoretical analysis and numerical results. through the MIMO channel [15].
It is shown that at medium signal-to-noise ratio, beam-nulling In this paper, we propose a new technique called “beam-
approaches the optimal water-filling scheme.
nulling” (BN). This scheme uses the same feedback bandwidth
I. I NTRODUCTION as beamforming. That is, only one eigenvector is fed back
to the transmitter. The beam-nulling transmitter is informed
It has been recognized that adaptive techniques, proposed with the weakest spatial subchannel and then sends signals
for single-input-single-output (SISO) channels [1] [2], can also over a generated spatial subspace orthogonal to the weakest
be applied to improve MIMO channel capacity [3]- [5]. The subchannel. Note that both transmitter and receiver should
ideal scenario in adaptive schemes is that the transmitter has know how to generate the same spatial subspace. Hence,
full knowledge of channel state information (CSI) which is the loss of channel capacity as compared to the optimal
fed back from the receiver. With such a perfect CSI feedback, water-filling scheme can be reduced. Although the transmitted
the original MIMO channel can be converted to multiple symbols are “precoded” according to the feedback, the beam-
uncoupled SISO channels via singular value decomposition nulling scheme is different from the other existing precoding
(SVD) at the transmitter and the receiver [3]. In other words, schemes with limited feedback channel, which are independent
the original MIMO channel can be decomposed into several or- of the instantaneous channel but the optimal precode depends
thogonal “spatial subchannels” with various propagation gains. on the instantaneous channel [16] [17].
To achieve better performance, various strategies to allocate
constrained power to these subchannels can be implemented II. C HANNEL M ODEL
depending on the availability of CSI at the transmitter [6]- [8]. In this study, the channel is assumed to be a Rayleigh flat
If the transmitter has full knowledge about channel matrix, fading channel with Nt transmit and Nr (Nr ≥ Nt ) receive
i.e., full CSI, the so-called “water-filling” (WF) principle is antennas. We denote the complex gain from transmit antenna
performed on each spatial subchannel to maximize the channel n to receiver antenna m by hmn and collect them to form
capacity [3]. This scheme is optimal in this case. Various WF- an Nr × Nt channel matrix H = [hmn ]. The channel is
based schemes have been proposed, such as [9]- [12]. For known perfectly at the receiver. The entries in H are assumed
the WF-based scheme, the feedback bandwidth for the full to be independently identically distributed (i.i.d.) symmetrical
CSI grows with respect to the number of transmit and receive complex Gaussian random variables with zero mean and unit
antennas and the performance is often very sensitive to channel variance.
estimation errors. The symbol vector at the Nt transmit antennas is denoted by
Various beamforming techniques for MIMO channels have x = [x1 , x2 , . . . , xNt ]T . According to information theory [5],
also been investigated extensively, which can mitigate the the optimal distribution of the transmitted symbols is Gaussian.
above disadvantages. In an adaptive beamforming scheme, Thus, the elements {xi } of x are assumed to be i.i.d. Gaussian
complex weights of the transmit antennas are fed back from variables with zero mean and unit variance, i.e., E(xi ) = 0
the receiver. If only one eigenvector can be fed back, eigen- and E|xi |2 = 1.
beamforming [7] is optimal. The eigen-beamforming scheme The singular-value decomposition of H can be written as
only applies to the strongest spatial subchannel but can achieve
H = UΛVH (1)
full diversity and high signal-to-noise ratio (SNR) [7]. Also, in
practice, the eigen-beamforming scheme has to cooperate with where U is an Nr ×Nr unitary matrix, Λ is an Nr ×Nt matrix
the other adaptive parameters to improve performance or/and with singular values {λi } on the diagonal and zeros off the
Ant-1
g1,1
diagonal, and V is an Nt ×Nt unitary matrix. For convenience, x1 Ant-1
g1
we assume λ1 ≥ λ2 . . . ≥ λNt , U = [u1 u2 . . . uNr ] and gNt,1