参数资料
型号: TPTDSCDMAWP
英文描述: Spatial Temporal Processing in TD-SCDMA
中文描述: 时间空间处理在TD - SCDMA
文件页数: 6/23页
文件大小: 415K
代理商: TPTDSCDMAWP
Spatial Temporal Processing in TD-SCDMA
Error! Reference source not found.
14
5.
Non-directional channel impulse response
We will estimate the non-directional channel impulse response first. The estimated response can
be used to estimate the DoA for each user and each path. Then based on the estimated DoA and
non-directional channel impulse response, the directional channel impulse response can be
derived. The estimation will be based on the midamble training sequence with length of W+L.,
where W is the maximum number of channel delay taps.
Suppose we have x = Ad + n, where n is a Nx1 Gaussian noise vector (it is not necessary to be
white), A is an Nxp known matrix, d is px1 signal vector and x is Nx1 observed signal, then a ML
estimator of d can be derived as
x
n
1
H
1
H
)
(
=
R
A
R
A
d
(20)
Actually, the estimator
d is an efficient MVU (Minimum Variance Unbiased estimator reaching
the Cramer-Rao Lower Bound) if the vector n is a complex Gaussian noise vector.
Now, we define Ka Lx1 column vectors:
em
(ka)
for ka=1…Ka.
em
(ka)
is the received signal
for the antenna ka based on the last L midamble training sequence. Stack Ka vectors together to
form the received L x Ka matrix
Em = [em
(1) , e
m
(2) ,…, e
m
(Ka) ]
Similarly, we define received noise vectors and LxKa matrix
Nm = [nm
(1) , n
m
(2) ,…, n
m
(Ka) ]
Then we have
Em= GH + Nm
where the matrix G is an L x KW observing matrix.
Moreover we have
em= vec{ Em }= vec{ GH } + vec{Nm} = vec{ GHI
(Ka) } + n
m
(21)
= (I
(Ka)
G)vec{H}+ n
m = (I
(Ka)
G)h+ n
m
where I
(Ka) is a K
a-by-Ka identity matrix, and
is the Kronecker product operator.
Thus, from (8), we have
m
H
ka
m
H
ka
e
R
G
I
G
I
R
G
I
h
1
)
(
1
)
(
1
)
(
)
(
)}
(
)
{(
=
(22)
Now we have to work out the matrix G and the noise covariance matrix Rm before the non-
directional channel impulse response vector h can be obtained.
We define the matrix G to be
G = [G
(1), G(2),…, G(K)]
(23)
where G
(k)
k=1…K is the LxW Toeplitz matrix of the midamble training sequence for the kth
user. It is clear that the matrix G is pre-defined since it is composed from the given midamble
training code for all K users.
For the noise covariance matrix Rm, we have
相关PDF资料
PDF描述
TPU1897 TRANSISTOR | JFET | N-CHANNEL | 30MA I(DSS) | TO-226AA
TPU1898 TRANSISTOR | JFET | N-CHANNEL | 15MA I(DSS) | TO-226AA
TPU1899 TRANSISTOR | JFET | N-CHANNEL | 8MA I(DSS) | TO-226AA
TPU290 TRANSISTOR | JFET | N-CHANNEL | 500MA I(DSS) | TO-92
TPU291 TRANSISTOR | JFET | N-CHANNEL | 200MA I(DSS) | TO-92
相关代理商/技术参数
参数描述
TPTELTVWCC10 制造商:Pass & Seymour/Legrand 功能描述:CC ABS WP 1 RJ11 JACK 1 CATV F CON
TPTPAMAM166 制造商:PERKINELMER 制造商全称:PerkinElmer Optoelectronics 功能描述:Array Module
TPU100B 制造商:Bogen Communications 功能描述:Amplifier; 100 W (RMS); 70 Hz to 15 kHz 1 dB; ALC; Wall; 21 lbs. 制造商:Bogen Communications 功能描述:AMPLIFIER, TPU SERIES TELEPHONE PAGING; Leaded Process Compatible:No; Peak Reflow Compatible (260 C):No ;RoHS Compliant: No
TPU100B-W1 制造商:Black Box Corporation 功能描述:1 year warranty for TPU100B
TPU100B-W3 制造商:Black Box Corporation 功能描述:3 year warranty for TPU100B