Scribd
Upload
101 views11 pages

8-Channel Models Estimation

This document describes techniques for estimating channel characteristics from input-output data in a communication system. It discusses estimating the impulse response of a linear time-invariant channel using cross-correlation of the input and output signals. For a linear time-varying channel, the input-output data is divided into blocks, and the impulse response is estimated within each block to determine how the channel varies over time and frequency due to effects like multipath and Doppler spread. Matlab examples are provided to estimate time and frequency spread from the channel impulse responses.

Uploaded by

Ladissa Harvidianti
Copyright
© Attribution Non-Commercial (BY-NC)
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
101 views11 pages

8-Channel Models Estimation

This document describes techniques for estimating channel characteristics from input-output data in a communication system. It discusses estimating the impulse response of a linear time-invariant channel using cross-correlation of the input and output signals. For a linear time-varying channel, the input-output data is divided into blocks, and the impulse response is estimated within each block to determine how the channel varies over time and frequency due to effects like multipath and Doppler spread. Matlab examples are provided to estimate time and frequency spread from the channel impulse responses.

Uploaded by

Ladissa Harvidianti
Copyright
© Attribution Non-Commercial (BY-NC)
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
You are on page 1/ 11

Channel Estimation from Data

1. Recall Impulse Response Identification from


Correlation
2. Estimation of Time Spread and Doppler Shift
3. Simulink/Matlab Example
4. Stanford University Interim (SUI) Channel Models
Estimation of Channel Characteristics from Input - Output data.
1. For Linear Time Invariant (LTI) systems:
] [n x
] [n y
] [n h
[ ] [ ]* [ ] [ ] [ ] y n h n x n h x n
+
=
= =

Excite the system with white noise and unit variance


{ } ] [ ] [ ] [ ] [
*
m m n x n x E m R
xx
o = =
m
and compute the crosscorrelation between input and output
{ }
{ }
*
*
[ ] [ ] [ ]
[ ] [ ] [ ] [ ] [ ] [ ]
yx
R m E y n x n m
h E x n x n m h m h m o
+ +
= =
=
= = =

In matlab:
x
y
] [n x
] [n y
?
1. Get data (same length for simplicity):
2. Compute crosscorrelation between input and output:
h=xcorr(x,y);
If x[n] is white noise, h[n] is the impulse response.
2. For a Linear Time Varying Channel:
Goal: estimate time and frequency spread.
Known:
1. Sampling frequency
2. Upper bound on max Doppler Frequency
s
F
max D
F
Multipath Rayleigh
Fading Channel
Rayleigh
Fading
] [n x
[ ] y n
[ , ] h n
[ ] [ , ] [ ] y n h n x n
+
=
=

The impulse
response changes
with time
1. Collect Data and partition in blocks of length
:
x
y

1/
MAX
s D
N T F <<
max D
s
F
F
N <<
X=reshape(x,N,length(x)/N);
Y=reshape(x,N,length(y)/N);
[ ]
N
X,Y =
0 = n N n = N n 2 =
N n =
B
N

N N n
B
=
Within each block the channel is
almost time invariant
2. Estimate impulse response in each block :
h(:,i)=xcorr(Y(:,i),X(:,i))/N;

h =
[
B
N
]
2 1 N
h =
[

]
B
N
2 1 N
Each row is an impulse response
taken at different times
t
s
F N / /
s
N F +
Take the transpose:
plot((-N+1:N-1)/Fs, abs(h(:,i)));

3. Compute Power Spectrum on each column of h (each row of h) , to determine


time variability of the channel (If the channel is Time Invariant all columns of h are
the same):
h =
[

H=fft(h);
S=H.*conj(H);
]

B
N
2 1 N

[
B
N
]
2 1 N
time
time
s
T t =
( )
s
t n NT =
S =
time
s
T t =
Freq.
s
B
F
F k
N N
=

4. Take the sum over rows for Doppler Spread and sum over columns for
Time Spread (fftshift each vector to have zero term (sec or Hz) in the
middle
s
F m t / =
B
s
N
N F
k f
) / (
=
St
Sf
Frequency Resolution:
Hz
N N
F
F
B
S
) length(sec data total
1
=

= A
Therefore if we want to a resolution in the doppler spread of (say) 1Hz,
we need to collect at least 1 sec of data.
Time Resolution:
S
F t / 1 = A
Example: % channel
Fs=10^6; % sampling freq. In Hz
P=[0,-2,-3]; % attenuations in dB
T=[0, 10, 15]*10^(-6); % time delays in sec
fd=70; %doppler shift in Hz
y
To Workspace1
x
To Workspace
Rectangul ar
QAM
Rectangul ar QAM
Modul ator
Baseband
Rayl ei gh
Fadi ng
Mul ti path Rayl ei gh
Fadi ng Channel
Bernoul l i
Bi nary
Bernoul l i Bi nary
Generator
test_scattering.mdl
Channel Output (Magnitude) with a QPSK Transmitted Signal:
t (sec)
0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 0.2
0
0.5
1
1.5
2
2.5
3
-1.5 -1 -0.5 0 0.5 1
x 10
-4
0
1
2
3
4
5
6
7
8
9
x 10
-3
Time Spread
time (sec)
-1000 -800 -600 -400 -200 0 200 400 600 800 1000
0
0.2
0.4
0.6
0.8
1
1.2
x 10
-3
Frequency Spread
frequency (Hz)
Time Spread
Hz 70 + Hz 70
Frequency Spread
sec 15
sum(S)/NB; sum of each column
sum(S)/(2N-1); ave. of each row

You might also like