Filtering Of Chaotic Signals And Its Applications In Chaotic Modulation Communication Systems

Chaotic signals that result from deterministic nonlinear systems are non-periodic and bounded. They are sensitive to initial conditions and look like "noise" in time domain. Because of the noise-like feature, chaotic signals usually occupy a wide bandwidth in frequency domain. Those characteristics mentioned above make it have potential applications in engineering practice. Since the self-synchronization phenomenon of chaotic systems was found, the research for communications with chaos has been greatly evoked in domestic and international academic societies, and many kinds of communication systems based on chaos have been proposed.Most researches on chaos-based communication systems are based on the assumption of a rather ideal communication environment, in which signals are transmitted without distortion or with only a moderate amount of additive noise. Besides the existence of random noise in real physics channel, the limited bandwidth impairs the amplitude and phase of the transmitted broadband chaotic signals, the multipath effect in wireless communication channel and the time-varying feature of channel can also distort the transmitted chaotic signals; in addition, the echo effect in communication systems has a serious degradation of quality of service of communication systems. These factors have impaired the realization of the proposed chaos-based communication system. Filtering is an important technique for combating these unwanted interferences in chaos-based communication systems.This thesis focuses on the topic of chaotic modulation communication systems, covering the analyzing for several adaptive filtering algorithms; proposing an improved adaptive filtering algorithm; filtering for noisy contaminated chaotic signals by combining the estimate theory; designing an adaptive demodulator and a blind equalizer, proposing an echo cancellation strategy based on adaptive filtering algorithms and the modeling technique for systematic parameters. The extended Kalman filter algorithm, the least mean squared and FIR adaptive filtering algorithms are used to demodulate noisy contaminated chaotic signals. The transmitted message signals are hidden in parameters of chaos generators, message signals to be considered in this thesis are constant signal, rectangle wave, ladder wave, sine wave, speech signal and image signal. After transferring the model of the filtering for chaotic signals going through fading communication scenario into the extended state space model, we realize the adaptive demodulation of the chaos-based communication systems by using the estimate theory combining with the adaptive filtering algorithms and autoregressive modeling technique for parameters. The results by computer simulation indicate that the adaptive demodulation algorithms by using the combined adaptive filtering and autoregressive modeling can effectively reduce channel's noise and recover message signals. The computing complexity and demodulation performance of several adaptive filtering methods are analyzed. Hereafter the adaptive filtering algorithm is also applied to the blind channel equalization for canceling multipath fading effect in chaos-based communication systems. The simulation results for those channels with constant and time-varying coefficients demonstrate that the algorithm can realize the equalization tasks. An improved adaptive filtering algorithm, in which a function related to the filter's input and the systematic vectors is used, is employed to adjust the step size. The algorithm can effectively reduce noise and preferably combat interferences. The algorithm is also applied to the echo cancellation in chaotic speech-modulation communication systems. The results by computer simulation indicate that the algorithm can realize single-channel echo cancellation.