Özet:

One of the most important requirements for speech channel is to provide information security. Speech scrambling is one of the preferred methods for secure speech communications over the narrowband channels. That confirms the relevance of works focused on improvement and investigation of operation features of speech scrambling devices. The article shows the advantages of using scramblers in narrowband voice communication systems and presents a detailed analysis of strength to hacking of communication channels secured by frequency-domain scrambling technique. The investigation of strength to hacking is based on the developed digital signal processing algorithm which realizes the band scrambler. The developed algorithm does not require synchronization between the scrambler and the descrambler which is achieved by using of sliding window FFT technique. Presented experimental results show that the lowest level of the scrambled signal residual intelligibility can be achieved when low frequency bands of speech spectrum are permuted with high frequency ones. Also the article shows that if amount of bands is less than 32 the scrambled signal can be hacked by a simple frequency spectrum inversion in the entire frequency band of the signal. During scrambling with amount of bands greater than 32, there are variants of band permutation that ensure the level of residual intelligibility near 10‑20 % and that is secured against hacking by the spectrum inversion. For a real-time speech communication system, the maximum number of scrambler bands is limited by two factors: the delay time for the scrambling-descrambling process and inserted distortions. It was shown that the maximum amount of bands should be less or equal to 256 for scrambling time delay of 128 ms, communication channel bandwidth of 4 kHz and the level of distortions that causes reduction of intelligibility not lower than 90 %. The presented information can be used in development of frequency-domain scramblers and for selection scrambler keys, which are optimal by the criteria of strength to hacking, minimum of residual intelligibility and level of inserted distortions. Author Biographies Vitalii Kukush, Kharkiv National University of Radio Electronics Nauky ave., 14, Kharkiv, Ukraine, 61166 PhD, Associate Professor Department of computer radio engineering and systems of technical security of information

Anahtar Kelimeler: