Application Of Computational Intelligence Methods In Chaos-based Communication Systems

Recently, researchers from various fields of science and engineering have shown increasing interest in the application of chaos to communications. In this dissertation, we mainly pay attention to the key techniques for chaos-based communication systems. The main contents of this dissertation include: (1) Symbols synchronization using fractional delay filter for chaos-based communication systems; (2) Chaotic sequences in fractional order dynamic systems; (3)Adaptive equalization for fading channels of chaos-based communication systems; and (4) Noise reduction to chaos-based communications system;The main Originality in this paper can be summarized as follows:1. Study on symbols synchronization using fractional delay filter for chaos-based communication systems.In DCSK, every transmitted symbol is represented by two sets of chaotic signal samples sent in two slots. The first sample set serves as the reference while the second one carries the data. The judge for symbols transmitted is based on the output of the correlator which the reference sample and the corresponding data sample are correlated. Thus, the receiver must be synchronized to the symbols of the incoming data signal. For this application, it may be desirable to delay a signal by a fractional multiples of the sampling period. A neural network method is proposed to designing variable-fractional-delay finite-impulse response filters used for approximation to ideal fractional delay filters. The proposed method has improved performance of the filters at a slightly increased computation cost while being used in chaotic communication systems.2. Study on chaotic sequences in fractional order dynamic systems.We numerically investigate chaotic behavior in noninteger-order cellular neural networks, which give good results in generating complex dynamics. Linear transfer function approximations of the fractional integrator block are calculated based on frequency domain arguments. Numerical simulations are used to demonstrate that the hyperchaotic attractors are obtained with system order 3.95. The complex dynamics of fractional order CNN provide a new signal source for secure communication.