Static and differential precoding codebook for MIMO systems
View Patent ↗Systems and methods are disclosed to generate a codebook for channel state information by generating a random codebook; partitioning channel state information into a set of nearest neighbors for each codebook entry based on a distance metric; and updating the codebook by finding a centroid for each partition.
1. A method to generate a codebook for channel state information in a wireless system having multiple input/multiple output (MIMO) antennas, comprising:
a. generating a random codebook;
b. partitioning channel state information into a set of nearest neighbors for each codebook entry based on a distance metric; and
c. updating the codebook by finding a centroid for each partition
wherein the centroid for the k th partition comprises
q
i
2
(
2
)
=
arg
min
q
∈
ℂ
(
t
-
1
)
×
1
,
q
*
q
=
1
∑
i
1
=
1
K
1
𝔼
[
d
p
(
[
q
i
1
(
1
)
,
Ψ
(
q
i
1
(
1
)
-
e
t
)
[
0
q
]
]
,
V
(
2
)
)
ℐ
(
V
(
2
)
∈
v
i
1
,
i
2
(
2
)
)
]
.
2. The method of claim 1 , comprising optimizing the codebook.
3. The method of claim 1 , wherein the codebook Q comprises {Q 1 , Q 2 , . . . , Q 2 B }, comprising determining an optimal partitioning satisfying
V k ={VεC t×r :d p ( V,Q k )≦ d p ( V,Q j ),∀ j≠k}.
4. The method of claim 1 , comprising determining the centroid by solving for the k th partition
Q
k
=
argmin
J
(
Q
)
=
arg
min
Q
∈
C
txr
,
Q
*
Q
=
I
E
⌊
d
p
(
Q
,
V
)
|
V
∈
V
k
⌋
.
5. The method of claim 1 , comprising determining an optimal partitioning satisfying
V k (1) ={vεC t×1 :d p ( v,q k (1) )≦ d p ( v,q j (1) ),∀ j≠k}.
6. The method of claim 1 , comprising determining the centroid for the k th partition by
q
k
(
1
)
=
arg
min
q
∈
c
tx1
,
q
*
q
=
1
E
[
d
p
(
q
,
v
)
|
v
∈
V
k
(
1
)
]
.
7. The method of claim 1 , comprising generating a set of doubly indexed partitions.
8. The method of claim 7 , wherein each partition is identified by two vectors in the first and the second vector codebooks Q (1) and Q (2) .
9. A method to generate a codebook for channel state information in a wireless system having multiple input/multiple output (MIMO) antennas, comprising:
a. generating a random codebook;
b. partitioning channel state information into a set of nearest neighbors for each codebook entry based on a distance metric;
c. updating the codebook by finding a centroid for each partition; and
d. determining an optimal partitioning satisfying
v
i
1
,
i
2
(
2
)
=
[
V
(
2
)
=
[
v
(
1
)
v
(
2
)
]
∈
ℂ
t
×
2
:
d
p
(
V
(
2
)
,
[
q
i
1
(
1
)
,
Ψ
(
q
i
1
(
1
)
-
e
t
)
[
0
q
i
2
(
2
)
]
]
)
≤
d
p
(
V
(
2
)
,
[
q
j
1
(
1
)
,
Ψ
(
q
j
1
(
1
)
-
e
t
)
[
0
q
j
2
(
2
)
]
]
)
,
∀
(
j
1
,
j
2
)
≠
(
i
1
,
i
2
)
}
.
10. The method of claim 1 , wherein the codebook design is static.
11. The method of claim 3 , wherein the codebook is a rank specific codebook.
12. The method of claim 1 , wherein the codebook is a nested vector codebook.
13. The method of claim 1 , wherein the codebook comprises a differential codebook.
14. The method of claim 13 , wherein the differential codebook provides feedbacks only on variations in the channel.
15. The method of claim 1 , comprising applying Lloyd-Max technique.
16. A method to generate a codebook for channel state information in a wireless system having multiple input/multiple output (MIMO) antennas, comprising:
a. generating a random codebook;
b. partitioning channel state information into a set of nearest neighbors for each codebook entry based on a distance metric;
c. updating the codebook by finding a centroid for each partition; and
d. determining an optimal partitioning by satisfying:
V k ={( {circumflex over (V)},V )ε C t×r ×C t×r : ∥O k {circumflex over (V)}−V∥ F 2 ≦∥O j {circumflex over (V)}−V∥ F 2 ∀j≠k}.
17. A method to generate a codebook for channel state information in a wireless system having multiple input/multiple output (MIMO) antennas, comprising:
a. generating a random codebook;
b. partitioning channel state information into a set of nearest neighbors for each codebook entry based on a distance metric;
c. updating the codebook by finding a centroid for each partition; and
d. maximizing:
E
[
ℜ
etr
{
V
V
^
*
O
*
}
|
(
V
^
,
V
)
∈
V
k
]
=
ℜ
etr
{
E
[
V
V
^
|
(
V
^
,
V
)
∈
V
k
]
O
*
}
=
ℜ
etr
{
Θ
∑
Φ
*
O
*
}
=
ℜ
etr
{
∑
Φ
*
O
*
Θ
}
=
ℜ
e
∑
i
=
1
M
σ
i
s
ii
where ΘΣΦ*=SVD(E└V{circumflex over (V)}*∥({circumflex over (V)},V)εV k ┘) and T=[s ij ]=Φ*O*Θ is a unitary matrix.
18. A wireless system, comprising
a. one or more transmitters;
b. one or more receivers each having multiple input/multiple output (MIMO) antennas adapted to receive data from the one or more transmitters, and
c. a codebook formed by generating a random codebook; partitioning channel state information into a set of nearest neighbors for each codebook entry based on a distance metric; and updating the codebook by finding a centroid for each partition
wherein the centroid for the k th partition comprises
q
i
2
(
2
)
=
arg
min
q
∈
ℂ
(
t
-
1
)
×
1
,
q
*
q
=
1
∑
i
1
=
1
K
1
𝔼
[
d
p
(
[
q
i
1
(
1
)
,
Ψ
(
q
i
1
(
1
)
-
e
t
)
[
0
q
]
]
,
V
(
2
)
)
ℐ
(
V
(
2
)
∈
v
i
1
,
i
2
(
2
)
)
]
.
19. The system of claim 18 , wherein the codebook is generated using a Lloyd-Max technique.