Performance Enhancement of Speech Scrambling Techniques Based on Many Chaotic Signals

Abstract

scrambling method structure consists of three main components: scrambling, noisy channel and descrambling. The scrambling and descrambling processes. On the basis of chaotic signals, a speech scrambling technique is an essential work to upgrade and improve data protection in secure communication applications, especially military applications, and the main idea is immunity against eavesdropper to achieve security. The aim of uses the key encrypted in the Quadrature Amplitude Modulation-Orthogonal Frequency Division Multiplexing QAM OFDM system to improve performance of the Bits Error Rates BER on the receiving side that is robust against the effects of the Additive White Gaussian Noise AWGN. This paper has been published a speech authentication model using five separate chaotic maps / flows: Logistic maps LM, Baker maps BM, Henon map HM, Rössler system RÖS and Lorenz system LS based on 16-64 QAM_OFDM technique. The evaluation measures included six different tests: Segmental spectral signal to ratio of noise SSSNR, Log spectral distance dLOG, Linear predict coding distance dLPC, cepstral distance dCD, Frequency Weighted log Spectral distance dFWLOG and Bit error rate BER. Objective tests to measure the residual intelligibility R.I. of speech between the original and scrambling/ descrambling signal for proposed models: dLOG= 14.5415, dLPC= 0.9741, dCD= 8.8503, SSSNR= −26.506, and dFWLOG= 20.9976, While the BER between the original signal and recovered signal at 10 dB about 10-03, Which is big enough to protect the brute force attack encryption signal. We conducted a series of unpredictable map / flow randomness tests: Lyapunov exponent test and Auto correlation.