Digital Signal Processing Reference
In-Depth Information
B
2
D
23
⊥
D
22
.
Design
precoder
to
make
D
14
⊥
D
11
,
D
14
⊥
D
12
,
D
24
⊥
D
21
,
D
23
.
2. At time slot 2, design precoder
B
1
D
24
⊥
D
22
,
D
24
⊥
. Design precoder
A
1
to make
D
12
U
G
1
(
1
)
D
12
. Design precoder
B
2
to make
D
14
⊥
to make
D
11
⊥
D
11
,
D
14
⊥
D
12
,
D
24
⊥
D
21
,
A
2
D
24
⊥
D
22
. Design precoder
to make
D
13
⊥
D
12
,
D
13
⊥
D
11
,
D
23
⊥
D
21
,
D
24
.
3. At time slot 3, design precoder
A
2
D
23
⊥
D
22
,
D
23
⊥
. Design precoder
B
2
to make
D
23
U
H
2
(
1
)
D
23
. Design precoder
A
1
to make
D
11
⊥
to make
D
24
⊥
D
13
,
D
11
⊥
D
14
,
D
21
⊥
D
23
,
B
1
D
21
⊥
D
24
.
Design
precoder
to
make
D
12
⊥
D
13
,
D
12
⊥
D
14
,
D
12
⊥
D
11
,
D
24
.
4. At time slot 4, design precoder
B
2
D
22
⊥
D
23
,
D
22
⊥
. Design precoder
A
2
to make
D
24
U
G
2
(
1
)
D
24
. Design precoder
B
1
to make
D
12
⊥
to make
D
23
⊥
D
13
,
D
12
⊥
D
14
,
D
22
⊥
D
23
,
A
1
D
22
⊥
D
24
. Design precoder
to make
D
11
⊥
D
12
,
D
11
⊥
D
13
,
D
11
⊥
D
14
,
D
24
.
Note that the design method at each time slot is similar. The key is that we change
the design order for
C
1
,
S
1
,
C
2
,
S
2
at different time slots. At time slot 1, we should
design precoder for
C
1
, then for
S
1
, then for
C
2
, finally for
S
2
. At time slot 2, we
should design precoder for
S
1
, then for
C
1
, then for
S
2
, finally for
C
2
. At time slot
3, we should design precoder for
C
2
, then for
S
2
, then for
C
1
, finally for
S
1
.Attime
slot 4, we should design precoder for
S
2
, then for
C
2
, then for
S
1
, finally for
C
1
.
In what follows, we prove that our proposed scheme can provide full diversity
for each codeword. We only provide the proof for codewords
c
11
,
c
12
,
c
13
,
c
14
.The
proof for other codewords is similar. The diversity is defined as
D
21
⊥
D
23
,
D
21
⊥
log
P
e
log
d
=−
lim
ρ
→∞
(5.50)
ρ
where
ρ
denotes the SNR and
P
e
represents the probability of error. If we let
⎛
⎞
⎛
⎞
⎛
⎞
e
1
e
2
e
3
e
4
c
11
c
12
c
13
c
14
c
11
⎝
⎠
⎝
⎠
−
⎝
⎠
c
12
e
=
=
denote the error vector, based on Eq. (
5.42
),
c
13
c
14
the pairwise error probability (PEP) for
c
11
,
c
12
,
c
13
,
c
14
can be written as [
2
]
⎛
⎞
2
F
ρ
||
H
1
e
||
⎝
⎠
P
(
c
→
c
|
H
1
)
=
Q
4
⎛
⎞
exp
e
†
†
H
1
e
e
†
†
H
1
e
ρ
(
H
1
)
−
ρ
(
H
1
)
⎝
⎠
≤
=
Q
4
4
exp
−
ρ
4
=
(5.51)
Search WWH ::
Custom Search