Modelling of WCDMA channels for channel estimation in smart antenna systems

1
Modelling of WCDMA channels for channel
estimation in smart antenna systems

• June 18th 2002
• Gerard Rauwerda

2
Presentation overview

• Wireless communication channels
• Modelling of wireless communication channels
• Simulation results
• Conclusions recommendations

3
Wireless communication channels
Multiple paths due to scattering

• angle spread
• time-delay spread

4
Wireless communication channels
Attenuation of signal power

• path loss
• large-scale fading
• dynamics of environment
• small-scale fading
• incoherent superposition

5
Modelling of wireless channels

• Cell type (Macro- / Micro- / Pico-)
• external parameters height, frequency
• Radio Environment (BU / TU / RA / HT)
• global parameters power profiles
• Propagation scenario
• local parameters multi-path component
  clusters (of scatterers)

6
6
7
Modelling of wireless channels
Large-scale fading

• dynamics of the environment
• birth / death of clusters
• visibility areas

8
Modelling of wireless channels
Small-scale fading

• incoherent superposition of radio waves
• small-scale displacement

9
MIMO system

• Multiple transmit antennas
• Multiple receive antennas

N?M MIMO system min(N,M) independent parallel
channels ? pipes
10
MIMO system
Communication systems are described by
Singular Value Decomposition looks for
independent paths. The squared Singular
Valuesare the power gains of these paths.
11
Simulation results - SVD

• Channel Impulse Responses in different
  environments (BU / TU / RA / HT)
• 4 antennas at Base Station and Mobile Station
• Analyse MIMO channels throughSingular Value
  Decomposition
• wideband
• narrowband

12
Results - wideband SVD

• Consider power in a broad frequency band
• 100 channel realisations

13
Results - wideband SVD
13
14
Results - narrowband SVD

• Consider power in a narrow frequency band
• 100 channel realisations

15
Results - narrowband SVD
15
16
Results - SVD

• wideband MIMO systems
• 1 dominant pipe
• remaining pipes at least 20 dB worse
• narrowband MIMO systems
• 1 dominant pipe
• gain of 2nd pipe about 5 dB worse
• gain of 3th pipe about 10 dB worse
• Spacing between antenna elements has most impact
  on narrowband Singular Values

17
Simulation results - capacity

• Squared Singular Values denote gains of the pipes
• Determine capacity of MIMO systems withm-QAM
  modulation and waterfilling
• Bit Error Rate of 1 and 0.1

18
Narrowbandcapacityin BU
18
19
Narrowbandcapacityfor ½ ?
19
20
Narrowband widebandcapacity
20
21
Results - capacity

• 4 bps / Hz capacity gain for 3 dB SNR increase
• Best performance in Bad Urban
• Dependency on antenna spacing
• ½ ? at Mobile Station
• gt 5 ? at Base Station
• Capacity gain wideband 200 - 270
• Capacity gain narrowband 275 - 425

22
Conclusions

• Independent parallel channels ? pipes
• ratio of gains dependent on antenna spacing
• Narrowband system better than wideband system (
  100 - 150 )
• Best performance in Bad Urban
• Dependency on antenna spacing
• ½ ? at Mobile Station
• gt 5 ? at Base Station

23
Recommendations

• Eigen Value Decomposition instead of Singular
  Value Decomposition
• Fading correlation

24
Questions