Skip to main content

Table 10 Confusion Matrix at SNR = 4 dB.

From: Automatic Modulation Recognition Using Wavelet Transform and Neural Networks in Wireless Systems

Channel model

Full-class recognition

GSM/EDGE channel models

 

Typical case for rural area (RA100), 6 taps

99.7%

Typical case for hilly terrain (HT80), 12 taps

99.1%

Typical case for hilly terrain (HT80), 6 taps

98.3%

Typical case for urban area (TU60), 12 taps

99.2%

Typical case for urban area (TU60), 6 taps

97.7%

Typical case for very small cells (TIx)

99.2%

COST 207 channel models

 

Rural Area (RA100), 6 taps

99.2%

Typical Urban (TU60), 12 taps

99.7%

Bad Urban (BU60), 12 taps

97.9%

Hilly Terrain (HT80), 12 taps

99.3%

ITU-R 3G channel models

 

Indoor office (IA5)

98.4%

Outdoor to indoor and pedestrian (PA10)

99.7%

Vehicular - high antenna (VA100)

97.7%

Satellite, LOS (SA100LOS)

99.3%

Satellite, NLOS (SA100NLOS)

99.1%

  1. Algorithm performance of full-class recognition for SNR = 4 dB over different fading channels.