عنوان البحث(Papers / Research Title)
Speech Scrambling Based on Wavelet Transform
الناشر \ المحرر \ الكاتب (Author / Editor / Publisher)
ستار بدر سدخان المالكي
Citation Information
ستار,بدر,سدخان,المالكي ,Speech Scrambling Based on Wavelet Transform , Time 15/02/2013 17:19:26 : كلية تكنولوجيا المعلومات
وصف الابستركت (Abstract)
Speech Scrambling Based on Wavelet Transform
الوصف الكامل (Full Abstract)
introduction
the increased interest in analog speech scrambling techniques are due to the increased visibility and publicity given to the vulnerability of communication systems to eavesdropingping of unauthorized remote access (gersho & steele, 1984). in wireless communications, including high frequency (h.f) and satellite communications, it is almost impossible to prevent unauthorized people from eavesdropingping unless speech scramblers may be used to protect privacy. among speech scramblers, analog scramblers are attractive and wide applicable. the conventional analog scramblers manipulate speech signal in the frequency or time domain or both. a typical frequency domain scrambler is the band splitting scrambler, which breaks the speech signal into several sub bands and permutes them. a typical time domain scrambler is the time division scrambler, which breaks the speech signal into short time segments and permutes them within a block of several segments(sakurai et al., 1984). these conventional analog scramblers cannot provide sufficient security against cryptanalysis because the number of permutable elements in these scramblers is not large enough to provide an adequate number of different permutations due to hardware limitation and processing delays.
to strengthen security, a two-dimensional scrambler which manipulates the speech signal both in the frequency domain and in the time domain was proposed. regarding other types of scramblers, which can attain a high degree of security, the transform domain scrambler was proposed.in 1979, wyner proposed a method, in which the orthogonal transform called a prolate spheroidal transform (psd) was executed on a set of the sampled speech signal. a mathematical basis for using both band splitting and time division, at the same time was presented by f. pichler in 1983. he showed how an operation which realizes band splitting and time division can be designed, and pointed out that such an operation can be realized by a fast algorithm. the mathematical background is the theory of group-character for finite abelian groups and the theory of the general fast fourier transform (gfft) (pichler, 1983). also in 1984 lin-shan et. al., presented frequency domain scrambling algorithm, which is an extension of the discrete fourier transform (dft) scrambler previously proposed. the use of short-time fourier analysis and filter bank techniques lead to the special feature that the original speech could be correctly recovered while the frame synchronization is completely unnecessary. in 1990 sridharan et. al., presented a comparison among five discrete orthogonal transforms in speech encryption systems. the results of the research showed that the discrete cosine transform (dct) and the discrete prolate spheroidal transform (dpst) could be used in narrow band systems. the karhunen loeve transform (klt) and the discrete hadamard transform (dht) were more suitable where wider bandwidth was available. the dct turned out to be the best transform with respect to residual intelligibility of the encryption speech and recovered speech quality. the dft produced results which were inferior to the dct. the dct implementation would also offer speed advantage over the fft (sridharan et al., 1990).
in 2000, an automated method for cryptanalysis of dft-based analog speech scramblers was presented by wen-whei and heng-iang, through statistical estimation treatments. in the proposed system, the cipher text only attack was formulated as a combinatorial optimization problem leading to a search for the most likely key estimate. for greater efficiency, they also explored the benefits of genetic algorithm to develop the method. simulation results indicated that the global explorative properties of genetic algorithms make them very effective at estimating the most likely permutation and by using this estimate significant amount of the intelligibility could be recovered from the cipher text following the attack on dft-based speech scramblers (whei & iang, 2000)a time-frequency scrambling algorithm based on wavelet packets was proposed by ajit s. b. bopardikar (1995) by using different wavelet packet filter banks, they added an extra level of security since the eavesdropingper had to choose the correct analysis filter bank, correctly rearrange the time-frequency segments, and choose the correct synthesis bank to get back the original speech signal. simulations performed with this algorithm give distance measures comparable to those obtained for the uniform filter bank based algorithm( bopardikar, 1995). in 2005, an analog speech scrambler which is based on wavelet transformation and permutation was proposed by sattar b. sadkhan and evaluating the scrambling efficiency through the calculation of distance measures, and takes the effect of the channel noise into consideration (sadkhan, et al., 2005). in 2007, a parallel structure of different wavelet transforms were applied for speech scrambling. the proposed structure provided a good results in comparison with the system implemented in 2005 (sadkhan, falah, 2007) .
تحميل الملف المرفق Download Attached File
|
|