Loading...
| Friend's email | |
| Your name | |
| Your email | |
| enter code | |
This page was sent successfuly
- Type of Document: M.Sc. Thesis
- Language: Farsi
- Document No: 45054 (05)
- University: Sharif University of Technology
- Department: Electrical Engineering
- Advisor(s): Behnia, Fereidoon
- Abstract:
- Nowadays, with the increased penetration of wireless mobile networks (e.g. GSM, 3G, etc.), voice channels are widely available ubiquitously. This availability and wide coverage encourages utilization of such channels as a valuable available infrastructure for point to point secure data communication. This can be done either by using steganography and hiding the secure message in the voice passing through the channel or by ciphering and transmitting it through the voice channel as a speech-like signal. The proposed method for data communication or steganography through voice channels should be robust against lossy compression techniques applied by the vocoder block present in such channels. Moreover, it is shown in this thesis that the special compression algorithm used in voice channels make them nonlinear channels with long memory. Lossy compression, nonlinearity and memoryfulness along with some other constraints make data communication through voice-dedicated channels a challenging problem. In this thesis, different data communication (modulation and coding) techniques and steganography methods have been proposed such that they achieve acceptable data and bit error rate performance while maintaining maximum robustness against voice channel non-idealities. For efficient data communication, a heuristic optimization algorithm is proposed to devise a codebook of human voice samples that maximizes data transmission rate through the vocoder channel. Also, repetition coding is introduced as the optimum channel coding technique for communication through vocoder channel according to special characteristics of its compression process. Also some lower and upper bounds on voice channel capacity are derived and calculated by simulations on voice channel simulator. It is seen that there exists a large gap between previously reported rates in the literature and the derived lower bound. Therefore, the obtained bound encourages additional research to introduce new communication schemes that fulfill the existing gap. Finally, the Gilbert-Elliot channel model is introduced to calculate the overall capacity of the voice channel taking into account the adaptive vocoder change according to channel's traffic.Also, the performance of the Spread Spectrum (SS) steganography technique against vocoder attack in time and frequency domains is investigated in the second part of this thesis. Performance improvements are achieved by applying the cover signal estimation techniques at the reciever by utilizing conventional and adaptive filters. Finally in this section, the cover speech signal is modeled by Generalized Gaussian Distribution (GGD) and the optimum detector for the proposed method is derived. Performance investigation of the optimum detector for the blind and semi-blind cases is proposed as directions for future research
- Keywords:
- Watermarking ; Channel Capacity ; Voice Channel ; Lossy Compression ; Speech-Like Signal ; Global System for Mobile Communication (GSM)
Digital Object List-
محتواي کتاب - view
Bookmark