Channel Estimation in OFDM TechniquesSystems.ppt

Communication

1. Channel Estimation in
OFDM Systems
Zhibin Wu
Yan Liu
Xiangpeng Jing

2. OUTLINE
OFDM System Introduction
Channel Estimation Techniques
Performance Evaluation
Conclusion

3. OFDM Overview
Divides high-speed serial information signal into multiple
lower-speed sub-signals:
Transmits simultaneously at different frequencies in parallel.
Modulation ( BPSK, PSK,QPSK,16QAM, …).
Pilot sub-carriers used to prevent frequency and phase shift
errors.
Usage of cyclic prefix for lower multi-path distortion
Controlled overlapping of bands in one channel
Max spectral efficiency (Nyquist rate)
Easy implementation using inverse FFTs
Easy time-freq. Synchronization
Modulate by switching between time and frequency domain

4. Introduction to OFDM Systems

5. Time-Frequency View

6. Some Assumptions
Usage of cyclic Prefix
Impulse response of the channel shorter
than Cyclic Prefix
Slow fading effects so that the channel is
time-invariant over the symbol interval
Rectangular Windowing of the transmitted
pulses
Perfect Synchronization of transmitter and
receiver
Additive, white, Gaussian channel noise

7. System Architecture

8. System Architecture (cont’d)
1.Input to time domain
2.Guard Interval
3.Channel
4.Guard Removal
5.Output to frequency
domain
6.Output
7.Channel Estimation  1,...,2,1,0  NnkXIDFTnx 
 






1,...,1,0,
1,...,1,,
Nnnx
NNnnNx
nx
gg
f nwnhnxy
ff
  1,...,1,0  Nnnyny
f   1,...,2,1,0  NknyDFTkY 
1,...,1,0 

Nk
kWkIkHkXkY
ICIAWGNChannel


1,...,1,0  Nk
kH
kY
kX
e
e
Estimated
Channel

9. Pilot for Channel Estimation
Time
Carriers
Time
Carriers
Comb Type:
Part of the sub-
carriers are always
reserved as pilot for
each symbol
Block Type:
All sub-carriers is
used as pilot in a
specific period

10. Block-type Channel Estimation
LS: Least Square Estimation 




















1
0
110
1
.
.
.
,...,,
N
N
LS
y
y
y
xxxdiagXwhere
yXh

11. Comb-type Estimation 








0,
1,...,1,.inf
lmpx
Lldata
lmLXkX
Nppilot signals uniformly inserted in X(k)
L=Number of Carriers/Np
xp(m) is the mthpilot carrier value
{Hp(k) k=0,1,…,Np} , channel at pilot sub-carriers
Xp input at the kth pilot sub-carrier
Yp output at the kth pilot sub-carrier
LS Estimate


1,...,1,0 
p
p
p
p
Nk
kX
kY
kH LMS Estimate
Xp(k)
LMS +
e(k)-
Yp(k)

12. Interpolation for Comb-type
Linear Interpolation
Second Order
Interpolation 
  
Ll
mH
L
l
mHmH
lmLHkH
ppp
ee



0
1  
  






















Nl
m
p
Hcm
p
Hcm
p
Hc
c
c
c
where
lmLHkH
ee
/
1
10
1
1
,
2
1
1
,11
0
,
2
1
1





13. Simulation Parameters
Parameter Specifications
FFT Size 64
Number of Carriers 64
Pilot Ratio 1/16
Guard Length 16
Guard Type Cyclic Extension
data rate of OFDM signal 1Mbps/sub-carrier
Signal Constellation 16QAM

14. System structure in MATLAB Simulation

15. OFDM Transmitter
OFDM Receiver

16. Received and Recovered Signals
Received signal phases are distorted by multi-path fading

17. Comb-LS Estimation
Combating multipath rayleigh fading with RLS adaptive
equalization
A detail simulation with MATLAB
20 multipath, random phase, and weibull distribution of
amplitutde
Symbol Error Rate

18. Comb-LS Estimation

19. Filter length .vs. Sample Rate
1.Keep the ratio of F/S, increase S
2.Keep S, increase F.
Observed Symbol error rate with F ( filter length ) and
S ( samples per symbol)

20. Conclusion
OFDM System Introduction
Block Type
Direct or Decision Feedback
Comb Type
LS or LMS estimation at pilot frequencies
Interpolation Techniques
Linear
Second Order
Time Domain
Modulation
BPSK,QPSK,16QAM,DQPSK
Some Results:
Comb Type performs better since it tracks fast fading
channels.
RLS algorithm vs. LMS algorithm