# US 2009/0270118 A1 · Frequency Band Extrapolation Beamforming Weight Computation for Wireless Communications

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2009/0270118 A1
- Document ID: US20090270118A1
- Status: Published application
- Kind: A1
- Application number: 12/109,394
- Filed: 2008-04-25
- Published: 2009-10-29
- Earliest claimed: 2008-04-25
- Assignees: Cisco Technology, Inc.
- Inventors: Hang Jin; Yanxin Na
- Web page: https://patentgrep.com/patent/US20090270118A1

## CPC

- H04B 7/0617: {for beam forming}
- H04L 27/2602: {Signal structure}
- H04B: TRANSMISSION
- H04L: TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION (arrangements common to telegraphic and telephonic communication H04M)

## Abstract

Techniques are provided to facilitate the computation of beamforming weights used by a first communication device when sending a transmission via a plurality of antennas to a second communication device where knowledge of the behavior of the channel between the first communication device and the second communication device is limited to a portion of a wide frequency band. A frequency extrapolation beamforming weight computation process is provided to derive knowledge in other frequency subbands based on information contained in the received transmission from the second communication device.

## Claims

### Claim 1

A method comprising:
at a first communication device, receiving a transmission at a plurality of antennas from a second communication device, wherein the transmission uses some but not all frequency subbands within a frequency band space;
recovering baseband signals carried in the frequency subbands in the received transmission;
computing a channel covariance matrix from the baseband signals in the frequency subbands with respect to the plurality of antennas of the first communication device;
computing uplink beamforming weights from the baseband signals in the frequency subband from the channel covariance matrix;
computing a spatial projection of the uplink beamforming weights;
computing frequency domain variations in the channel between the first communication device and the second communication device based on difference in time of arrival of signals from different direction of arrivals at the plurality of antennas of the first communication device;
computing a downlink spatial projection for a given frequency subband from the frequency domain variations and the spatial projection of the uplink beamforming weights; and
computing downlink beamforming weights from the downlink spatial projection, wherein the downlink beamforming weights are for use by the first communication device when beamforming a downlink transmission via the plurality of antennas of the first communication device to the second communication device.

### Claim 2 (depends on 1)

The method of claim 1, wherein the transmission from the second communication device comprises a plurality of subcarriers at different frequencies in each of a plurality of tiles, and wherein computing the channel covariance matrix comprises computing J instantaneous channel covariance matrices in the frequency domain as a group of matrices {Rin}n=1J, where J is the number of tiles, i1, i2, . . . , iJ are the index of baseband signals for subcarriers in J tiles respectively, and wherein computing the uplink beamforming weights comprises computing a group of beamforming weight vectors {Win}n=1J for the J tiles.

### Claim 3 (depends on 2)

The method of claim 2, wherein computing the channel covariance matrix for a tile comprises computing
R
i
n
=
1
12

∑
j

∑
q

[
S
j
,
q

(
1
)
S
j
,
q

(
2
)
⋮
S
j
,
q

(
M
)
]

[
S
j
,
q

(
1
)
S
j
,
q

(
2
)
⋮
S
j
,
q

(
M
)
]
H
,
where a maximum number of subcarriers is L, such that 1≦i1≦i2< . . . <ij≦L, and Sj,q (m) for 1≦m≦M denotes received frequency domain signals on the jth subcarrier, the qth symbol, and mth antenna of the first communication device and M is the number of antennas of the first communication device.

### Claim 4 (depends on 2)

The method of claim 2, wherein computing the uplink beamforming weights is based on columns of the channel covariance matrices.

### Claim 5 (depends on 4)

The method of claim 4, wherein computing the uplink beamforming weights comprises maximal ratio combining of signals received at the plurality of antennas to obtain a sum of columns of channel covariance matrix Rin as Uin=rin,1+ejθ2rin,2+ . . . +ejθMrin,M, where Rin=[rin,1 rin,2 . . . rin,M], and rin,m, for 1≦m≦M, is the mth column vector in Rin, ejθm={rin,m}Hrin,1/(abs ({rin,m}H rin,1)), for 2≦m≦M, is the phase rotation of the mth column related to the first column, and abs( ) stands for absolute operation.

### Claim 6 (depends on 5)

The method of claim 5, wherein computing the uplink beamforming weights further comprises normalizing the sum of the columns of the channel covariance matrices to produce a beamforming weight Win as Win=Uin/(norm (Uin)), where norm( ) is the Euclidean norm operation, thereby obtaining a group of beamforming weight vectors {Win}n=1J for the J tiles.

### Claim 7 (depends on 2)

The method of claim 2, wherein computing the uplink beamforming weights is based on rows of the channel covariance matrices.

### Claim 8 (depends on 7)

The method of claim 7, wherein computing the uplink beamforming weights comprises maximal ratio combining of signals received at the plurality of antennas to obtain a sum of rows of the channel covariance matrices Rin as Uin={{tilde over (r)}in,1+ejφ2{tilde over (r)}in,2+ . . . +ejφM{tilde over (r)}in,M}H, where
R
i
n
=
[
r
~
i
n
,
1
r
~
i
n
,
2
⋮
r
~
i
n
,
M
]
,
{tilde over (r)}in,m, for 1≦m≦M is the mth row vector in Rin, ejφm={tilde over (r)}in,1{{tilde over (r)}in,m}H/(abs({tilde over (r)}in,1{{tilde over (r)}in,m}H)), for 2≦m≦M, is the phase rotation of the mth row related to the first row, and abs( ) stands for absolute operation and Hp is the Hermitian operation.

### Claim 9 (depends on 8)

The method of claim 8, wherein computing the uplink beamforming weights further comprises normalizing the sum of the rows of the channel matrices to produce a beamforming weight Win as Win=Uin/(norm (Uin)), where norm( ) is the Euclidean norm operation, thereby obtaining a group of beamforming weight vectors {Win}n=1J for the J tiles.

### Claim 10 (depends on 2)

The method of claim 2, wherein computing the spatial projection of the uplink beamforming weights comprises computing a spatial projection gain vector Gin for the beamforming weight vectors {Win}n=1J of the J tiles to Nsubsector subsectors in a space in which the first and second communication devices reside is divided, where Gin=[gin,1 gin,2 . . . gin,Nsubsector]T, gin,k is a complex-valued spatial projection gain for the subsector k and the received uplink tile n, and T denotes the Transpose operation.

### Claim 11 (depends on 10)

The method of claim 10, wherein computing the spatial projection gain of the uplink beamforming weights comprises computing, for each beamforming weight vector Win in the group of the beamforming weights vectors {Win}n=1J, a Fast Fourier Transform operation with size Nsubsector, to obtain the spatial projection gain vector Gin=FFT(Win, Nsubsector) in Nsubsector subsectors, and combining the spatial projection gains in all J received tiles as
P
K
=
∑
n
=
1
J


g
i
n
,
k

2
,
for each subsector k, where 1≦k≦Nsubsector.

### Claim 12 (depends on 10)

The method of claim 10, wherein computing the spatial projection gain comprises computing an array response vector A(θ, λ) for the plurality of antennas of the first communication device A(θ, λ) for a direction of arrival θ and a subcarrier wavelength λ associated with the received uplink transmission; computing a projection matrix Ξ=[A(φ1,λ) A(φ2,λ) . . . A(φNsubsector, λ)], for an angle φk in the subsector, for 1≦k≦Nsubsector; for each beamforming weight vector Win in the group of the beamforming weights vectors {Win}n=1J, applying the projection matrix Ξ to beamforming weight vector Win, to obtain the spatial projection gain vector Gin in Nsubsector subsectors as Gin=ΞHWin, where H is Hermitian operation.

### Claim 13 (depends on 10)

The method of claim 10, wherein computing the spatial projection gain further comprises comparing the spatial projection gain Pk summed over all J received uplink tiles for a sectork with a threshold such that
g
i
n
,
k
=
{
0
if


P
k
<
P
threshold
g
i
n
,
k
if


P
k
≥
P
threshold
for all received uplink tiles (1≦n≦J) and for all sectors 1≦k≦Nsubsector.

### Claim 14 (depends on 10)

The method of claim 10, wherein computing frequency domain variations comprises determining which subsector has the greatest power, denoted subsector kmax, where
k
max
=
arg


max
k

P
k
;
computing a vector of the relative spatial projection gain {tilde over (G)}in as
G
~
i
n
=

G
i
n
g
i
n
,
k
max
=

[
g
i
n
,
1
g
i
n
,
2
…
g
i
n
,
N
subsector
]
T
g
i
n
,
k
max
=

[
g
~
i
n
,
1
g
~
i
n
,
2
…
g
~
i
n
,
N
subsector
]
T
,
where T denotes for Transpose operation; and computing a time difference of arrival of signals at from subsectors in which |{tilde over (g)}in,k|>0 and the arrival of signals from the subsector kmax with the maximum power to produce a group of time difference values {αk} for all subsectors with |{tilde over (g)}in,k|>0.

### Claim 15 (depends on 14)

The method of claim 14, and further comprising computing the spatial projection gain for a frequency band spanned by Y subcarriers yl, such that the spatial projection gain for each subcarrier yl is {tilde over (G)}yl=[{tilde over (g)}yl,1 {tilde over (g)}yl,2 . . . {tilde over (g)}yl,Nsubsector]T, where T denotes the Transpose operation, and 1≦l≦Y.

### Claim 16 (depends on 15)

The method of claim 15, wherein computing the downlink spatial projection comprises computing, for a given subcarrier y, an index of a closest received tile to the subcarrier y,
i
closest
=
arg


min
i
n


y
-
i
n

,
where 1≦n≦J; and for kε[1, Nsubsector], applying the time difference αk using a Fast Fourier Transform size L, to produce the spatial projection gain for the frequency subcarrier y as
g
~
y
,
k
=
{
g
~
i
closest
,
k


j


2

πα
k

(
y
-
i
closest
)
/
L
if



g
~
i
closest
,
k

>
0
0
if



g
~
i
closest
,
k

=
0
where 2παk (y−iclosest)/L is a phase rotation to the spatial projection gain of the closest received tile; and computing the spatial projection gain for a frequency band spanned by Y subcarrier yl, such that the spatial projection gain for each frequency subcarrier yl is {tilde over (G)}yl=[{tilde over (g)}y1,1 {tilde over (g)}yl,2 . . . {tilde over (g)}yl,Nsubsector]T, where T denotes the Transpose operation, and 1≦l≦Y.

### Claim 17 (depends on 15)

The method of claim 15, wherein computing the downlink spatial projection comprises computing a vector Ωk as Ωk=ej2παk[i1 i2 . . . iJ]/L using the group of time difference values {αk}between the kth subsector and the kmax subsector, for all k with |{tilde over (g)}in,k|>0; computing a parameter hk=[{tilde over (g)}il,k {tilde over (g)}i2,k . . . {tilde over (g)}iJ,k] ΩkH from the vector Ωk and the spatial projection gain in J received uplink tiles {tilde over (g)}in,k, for 1≦n≦J, where H denotes the Hermitian operation; updating the spatial projection gain in the J uplink tiles using the parameters {hk} such that
g
~
i
n
,
k
=
{
h
k


j2πα
k

i
n
/
L
if



g
~
i
n
,
k

>
0
0
if



g
~
i
n
,
k

=
0
;
computing the spatial projection gain for subcarrier index y as
g
~
y
,
k
=
{
h
k


j2πα
k

y
/
L
if



g
~
i
1
,
k

>
0
0
if



g
~
i
1
,
k

=
0
,
where 2παky/L is the phase rotation based on the hk parameter of main path in kth subsector; and computing the spatial projection gain for a frequency band spanned by Y subcarriers yl, such that the spatial projection gain for each frequency subcarrier yl is {tilde over (G)}yl=[{tilde over (g)}yl,1 {tilde over (g)}yl,2 . . . {tilde over (g)}yl,Nsubsector]T, where T denotes the Transpose operation, and 1≦l≦Y.

### Claim 18 (depends on 14)

The method of claim 14, wherein computing the downlink beamforming weights comprises computing an Inverse Fast Fourier Transform of the estimated spatial projection vector {tilde over (G)}yl in the group of the estimated spatial projection vectors {{tilde over (G)}yl}l=1Y, to obtain a vector of the inversion of the spatial projection as {tilde over (W)}yl=IFFT ({tilde over (G)}yl, Nsubsector)=[{tilde over (w)}yl,1 {tilde over (w)}yl,2 . . . {tilde over (w)}yl,Nsubsector] and where the downlink beamforming weights WY in a frequency band spanned by Y subcarriers yl are Wyl=[{tilde over (w)}yl,1 {tilde over (w)}yl,2 . . . {tilde over (w)}yl,M]T/norm{[{tilde over (w)}yl,1 {tilde over (w)}yl,2 . . . {tilde over (w)}yl,M]T} for 1≦l≦Y.

### Claim 19 (depends on 12)

The method of claim 12, and computing the spatial projection gain in a frequency band spanned by Y subcarriers yl, such that the spatial projection gain for each frequency subcarrier yl is {tilde over (G)}yl=[{tilde over (g)}yl,1 {tilde over (g)}yl,2 . . . {tilde over (g)}yl,Nsubsector]T, where T denotes the Transpose operation, and 1≦l≦Y.

### Claim 20 (depends on 14)

The method of claim 14, wherein computing the downlink beamforming weights comprises computing for each estimated spatial projection vector {tilde over (G)}y1 in the group of the estimated spatial projection vectors {{tilde over (G)}yl}l=1Y, {tilde over (W)}yl={ΞΞH}−1Ξ{tilde over (G)}yl=[{tilde over (w)}yl,1 {tilde over (w)}yl,2 . . . {tilde over (w)}yl,M]T, where T denotes the transpose operator and H denotes the Hermitian operation, and where the downlink beamforming weights Wyl in a frequency band spanned by Y subcarriers yl are Wyl=[{tilde over (w)}yl,1 {tilde over (w)}yl,2 . . . {tilde over (w)}yl,M]T/norm {[{tilde over (w)}yl,1 {tilde over (w)}yl,2 . . . {tilde over (w)}yl,M]T} for 1≦l≦Y.

### Claim 21

An apparatus comprising:
a plurality of antennas;
a transmitter coupled to the plurality of antennas, wherein the transmitter is configured to apply beamforming weights to a transmit signal to generate corresponding ones of a plurality of weighted signals that are supplied to respective ones of the plurality of antennas;
a controller coupled to the transmitter configured to:
store data representing reception of a transmission sent by another communication device and received at the plurality of antennas, wherein the transmission uses some but not all frequency subbands within a frequency band space and wherein the data represents baseband signals carried in the frequency subbands of the received transmission;
compute a channel covariance matrix from data representing the baseband signals in the frequency subbands with respect to the plurality of antennas of the first communication device;
compute uplink beamforming weights from the baseband signals in the frequency subband from the channel covariance matrix;
compute a spatial projection of the uplink beamforming weights;
compute a downlink spatial projection for a given frequency subband from the spatial projection of the uplink beamforming weights; and
compute downlink beamforming weights from the downlink spatial projection, wherein the downlink beamforming weights are for use by the first communication device when beamforming a downlink transmission via the plurality of antennas of the first communication device to the second communication device.

### Claim 22 (depends on 21)

The apparatus of claim 21, wherein the controller is further configured to compute frequency domain variations in the channel between the first communication device and the second communication device based on difference in time of arrival of signals from different direction of arrivals at the plurality of antennas of the first communication device, and to compute the downlink spatial projection for a given frequency subband based on the spatial projection of the uplink beamforming weights and the frequency domain variations.

### Claim 23

Logic encoded in one or more tangible media for execution and when executed operable to:
store data representing reception of a transmission sent by a second communication device and received at a plurality of antennas of a first communication device, wherein the transmission uses some but not all frequency subbands within a frequency band space and wherein the data represents baseband signals carried in the frequency subbands of the received transmission;
compute a channel covariance matrix from data representing the baseband signals in the frequency subbands with respect to the plurality of antennas of the first communication device;
compute uplink beamforming weights from the baseband signals in the frequency subband from the channel covariance matrix;
compute a spatial projection of the uplink beamforming weights;
compute a downlink spatial projection for a given frequency subband from the frequency domain variations and the spatial projection of the uplink beamforming weights; and
compute downlink beamforming weights from the downlink spatial projection, wherein the downlink beamforming weights are for use by the first communication device when beamforming a downlink transmission via the plurality of antennas of the first communication device to the second communication device.

### Claim 24 (depends on 23)

The logic of claim 23, and further comprising logic configured to compute frequency domain variations in the channel between the first communication device and the second communication device based on difference in time of arrival of signals from different direction of arrivals at the plurality of antennas of the first communication device, and wherein the logic to compute the downlink spatial projection for a given frequency subband is configured to compute the downlink spatial projection based on the spatial projection of the uplink beamforming weights and the frequency domain variations.
