properties which remove
the noise, features extracted form details coefficients is
better than using features extracted from the approxima-
tion and wavelet coefficients and using two decomposi-
tion is better than one and three and bispectrum features
is better than higher order moment and cumulant features
and features extracted from the details discrete wavelet
transform coefficients using two decomposition levels.
We also note that, in the support vector machine classifiers,
the performance of the one-against-all classifier is better
than one-against-one and hierarchical support vector machine.
Finally, we note that, the performance of non clustering
techniques is better than clustering techniques but
clustering don’t need to train the classifier and it is easy
to implement and faster than non clustering techniques.
shows the features for the eight chirp signals when Ns is
32 samples. Figure 8 shows the contour plot of the eight
signals where Ns = 32 samples is used and the region R2 is
shown. The performance of the mutli-user chirp modula-
tion signals using multilayer perceptron neural network
and features F1, F2, F3, and F4 is shown in Figure 9.
Figure 10 shows the performance of the multi-user chirp
modulation signals using different classifiers and features
F4 is used. From these results, we note that using F4 as
features outperforms using F1, F2, and F3. Figure 11
shows the comparison between the performances of the
multi-user chirp modulation signals using multilayer per-
ceptron neural network classifier and the three different
features extraction methods using higher order moment
and cumulant (F4), using features extracted from the de-
tails discrete wavelet transform coefficients using two
decomposition levels, and bispectrum features (F4). From
this figure, we note that using bispectrum features is bet-
ter than using features extraction from higher order mo-
ment and cumulant (F4) and using features extracted
from the details discrete wavelet transform coefficients
using two decomposition levels at low signal to noise
ratio and the low signal power
From these results, we note that the performance using
Figure 8. Contour plot of the magnitude of the bispectrum of the eight signals on the bi-frequencies (f1, f2) and the region R2.
S. E. EL-KHAMY, H. A. ELSAYED
530
Figure 9. The performance of the multi-user chirp modulation signals using MLP classifier and bispectrum features.
Figure 10. The performance of the multi-user chirp modulation signals using different classifie r s and Bispectrum features F4.
Copyright © 2012 SciRes. IJCNS
S. E. EL-KHAMY, H. A. ELSAYED
Copyright © 2012 SciRes. IJCNS
531
Figure 11. Comparison between the performances of the multi-user chirp modulation signals using MLP classifier and dif-
ferent features ex traction methods.
7. Conclusions the classifier and it is easy to implement and faster than
non clustering techniques. Also, using features extracted
from WT get better performance than without WT be-
cause the WT have denoising properties which remove
the noise, features extracted form details coefficients is
better than using features extracted from the approxima-
tion and wavelet coefficients and using two decomposi-
tion is better than one and three and bispectrum features
is better than higher order moment and cumulant features
and features extracted from the details discrete wavelet
transform coefficients using two decomposition levels.
So this signal is type of UWB because it needs low signal
to noise ratio and then it is low signal power and also, we
deals with low power level spectrum signal, so this chirp
signal is type of spread spectrum signals.
In this paper, we presented classification of multi-user
chirp modulation signals using wavelet higher order sta-
tistics features and artificial intelligence techniques. In
this method, different types of classifiers are used and
different features extraction methods are used. We note
the dependence of the classifier performance on the clas-
sifier type, the classifier parameters, the features used,
the discrete wavelet coefficients, number of decomposi-
tion levels, the method of features extraction, and the
length of each segment.
Simulation results show that the performance of the
multilayer perceptron neural network classifier is better
than other classifiers such as maximum likelihood classi-
fier, support vector machine classifier, k nearest neighbor
classifiers, fuzzy c-means clustering, and fuzzy k-means
clustering because it take long time for training. In addi-
tion, the performance of the fuzzy c-means clustering is
better than the fuzzy k-means clustering for most the
SNR in case of using higher order moments and cumu-
lants as features. We also note that, in the support vector
machine classifiers, the performance of the one-against-
all classifier is better than one-against-one and hierarchi-
cal support vector machine. Finally, we note that, the
performance of non clustering techniques is better than
clustering techniques but clustering don’t need to train
REFERENCES
[1] C. R. C. M. da Silva, W. C. Headley, J. D. Reed and Y.
Zhao, “The Applications of Distributed Spectrum Sensing
and Available Resources Maps to Cognitive Radio Sys-
tems,” Proceedings of Information Theory and Applica-
tion Workshop, San Diegoq, 27 January-1 February 2008,
pp. 53-57.
[2] H. Guldemır and A. Sengur, “Comparison of Clustering
Algorithms for Analog Modulation Classification,” Ex-
pert Systems with Applications, An International Journal,
S. E. EL-KHAMY, H. A. ELSAYED
532
Vol. 30, No. 4, 2006, pp. 642-649.
[3] A. Ebrahimzadeh and M. Ebrahimzadeh, “An Expert Sys-
tem for Digital Signal Type Classification,” Journal of
Electrical Engineering, Vol. 58, No. 6, 2007, pp. 334-341.
[4] X. Zhou, Y. Wu and B. Yang, “Signal Classification Meth-
od Based on Support Vector Machine and High-Order
Cumulants,” Wireless Sensor Network, Vol. 2, No. 1, 2010,
pp. 48-52. doi:10.4236/wsn.2010.21007
[5] A. Ebrhimzadeh, H. Azimi and H. M. Naeemi, “Classifi-
cation of Communication Signals Using an Optimized
Classifier and Efficient Features,” The Arabian Journal
for Science and Engineering, Vol. 35, No. 1B, 2010, pp.
225-235.
[6] S. M. Zhou, J. Q. Gan and F. Sepulveda, “Classifying
Mental Tasks Based on Features of Higher-Order Statis-
tics from EEG Signals in Brain-Computer Interface,” In-
formation Sciences, Vol. 178, No. 6, 2008, pp. 1629-1640.
doi:10.1016/j.ins.2007.11.012
[7] C. K. Chua, V. Chandran, R. U. Acharya and L. C. Min,
“Cardiac Health Diagnosis Using Higher Order Spectra
and Support Vector Machine,” The Open Medical Infor-
matics Journal, Vol. 3, 2009, pp. 1-8.
doi:10.2174/1874431100903010001
[8] M. Kaur and A. S. Arora, “Unsupervised Analysis of Arr-
hythmias Using K-Means Clustering,” International Jour-
nal of Computer Science and Information Technologies,
Vol. 1, No. 5, 2010, pp. 417-419.
[9] M. Murugappan, M. Rizon, R. Nagarajan, S. Yaacob, I.
Zunaidi and D. Hazry, “EEG Feature Extraction for Clas-
sifying Emotions Using FCM and FKM,” Internat- ional
Journal of Computers and Communications, Vol. 1, 2007,
pp. 21-25.
[10] N. Ahmadi and R. Berangi, “Symbol Based Modulation
Classification Using Combination of Fuzzy Clustering
and Hierarchical Clustering,” Signal Processing: An In-
ternational Journal, Vol. 4, No. 2, 2010, pp. 123-137.
[11] M. Sayadi, L. Tlig and F. Fnaiech, “A New Texture Seg-
mentation Method Based on the Fuzzy C-Mean Algorithm
and Statistical Features,” Applied Mathematical Sciences,
Vol. 1, No. 60, 2007, pp. 2999-3007.
[12] P. Prakasam and M. Madheswaran, “Mary Shift Keying
Modulation Scheme Identification Algorithm Using Wave-
let Transform and Higher Order Statistical Moment,” Jour-
nal of Applied Science, Vol. 8, No. 1, 2008, pp. 112-119.
doi:10.3923/jas.2008.112.119
[13] R. M. Al-Makhlasawy, W. Hamouda, M. M. Abd Elnaby,
H. A. El-Khobby and F. E. Abd El-Samie, “Automatic
Modulation Recognition Using Cepestral Analysis and
Neural Networks in Wireless Systems,” 29th National
Radio Science Conference, Cairo, 10-12 April 2012.
[14] S. E. El-Khamy H. A. Elsayed and M. M. Rizk, “Classi-
fication of Multi-User Chirp Modulation Signals Using
Higher Order Cumulant Features and Four Types of Clas-
sifiers,” 28th National Radio Science Conference, Cairo,
Egypt, 26-28 April 2011.
[15] H. A. Elsayed, S. E. El-Khamy and M. R. Rizk, “Higher
Order Statistics Classification of Multi-User Chirp Modu-
lation Signals Using Clustering Techniques,” URSI GASS
2011, Istanbul, 13-20 August 2011.
[16] S. E. El-Khamy, H. A. Elsayed and M. R. Rizk, “Bispec-
trum Classification of Multi-User Chirp Modulation Sig-
nals Using Artificial Intelligent Techniques,” WorldComp
2011, Las Vegas, 18-21 July 2011.
[17] S. E. El-Khamy and M. B. Al-Ghoniemy, “The Wavelet
Transform: A Review and Application to Enhanced Data
Storage Reduction in Mismatched Filter Receivers,” Pro-
ceedings of the Thirteenth National Radio Science Con-
ference, Cairo, 19-21 March 1996.
[18] M. P. Fargues and G. Hatzichristos, “A Hierarchical Ap-
proach to the Classification of Digital Modulation Types
In Multipath Environments,” Conference Record of the
Thirty-Fifth Asilomar Conference on Signals, Systems and
Computers, Pacific Grove, 4-7 November 2001.
[19] C. L. Nikias and J. M. Mendel, “Signal Processing with
Higher Order Spectra,” IEEE Signal Processing Maga-
zine, Vol. 10, No. 3, 1993, pp. 10-37.
doi:10.1109/79.221324
[20] S. E. El-Khamy, “Matched Swept-Frequency Digital Mo-
dulation for Binary Signaling in Inhomogeneous Disper-
sive Media,” IEEE Transactions on Antennas and Pro-
pagation, Vol. 28, No. 1, 1980, pp. 29-35.
doi:10.1109/TAP.1980.1142283
[21] S. E. El-Khamy and S. E. Shaaban, “Matched Chirp Mo-
dulation: Detection and Performance in Dispersive Com-
munication Channels,” IEEE Transactions on Communi-
cations, Vol. 36, No. 4, 1988, pp. 506-509.
doi:10.1109/26.2777
[22] S. E. El-Khamy, S. E. Shaaban and E. A. Thabet, “Multi-
User Chirp Modulation Signals (M-CM) for Efficient
Multiple Access Communication Systems,” Proceedings
of the Thirteenth National Radio Science Conference,
Cairo, 19-21 March 1996.
[23] S. E. El-Khamy, S. E. Shaaban and E. A. Thabet, “Effic-
ient Multiple-Access Communications Using Multi-User
Chirp Modulation Signals,” IEEE 4th International Sym-
posium on Spread Spectrum Techniques and Applications
Proceedings, Mainz, 22-25 September 1996, pp. 1209-1213.
[24] S. E. El-Khamy, S. E. Shaaban and E. A. Thabet, “Fre-
quency Hopped-Chirp Modulation (FH-CM) for Multi-
User Signaling in Multipath Dispersive Media,” IEEE
Antennas and Propagation Society International Sympo-
sium, Orlando, 11-16 July 1999, pp. 396-399.
[25] C. L. Nikias and A. P. Petropulu, “Higher-Order Spectra
Analysis: A Nonlinear Signal Processing Framework,”
PTR Prentice-Hall, Englewood Cliffs, 1993.
[26] High Order Spectral Analysis Matlab Toolbox (9 February
2011).
http://www.mathworks.com/matlabcentral/fileexchange/3
013
[27] A. V. Rosti, “Statistical Methods in Modulation Classifi-
cation,” Master’s Thesis, Tampere University of Technol-
ogy, Tampere, 1999.
[28] H. Yoshioka, “A Fast Modulation Recognition Technique
Using Nearest Neighbor Rules with Optimized Threshold
for Modulation Classification in Rayleigh Fading Chan-
nels,” Proceedings of WPMC, Honolulu, 27-30 October
Copyright © 2012 SciRes. IJCNS
S. E. EL-KHAMY, H. A. ELSAYED
Copyright © 2012 SciRes. IJCNS
533
2002, pp. 1049-1052.
[29] C. Burges, “A Tutorial on Support Vector Machines for
Pattern Recognition,” Data Mining and Knowledge Dis-
covery, Vol. 2, No. 2, 1998, pp. 121-167.
doi:10.1023/A:1009715923555
[30] A. Swami and B. M. Sadler, “Hierarchical Digital Modu-
lation Classification Using Cumulants,” IEEE Transac-
tions on Communications, Vol. 48, No. 3, 2000, pp. 416-429.
doi:10.1109/26.837045