Study On Chaotic Frequency Hopping And Predistortion Technique Of Power Amplifier In Wireless Communication Systems

This thesis consists of two parts:Part One of this thesis addresses some achievements on developing chaotic frequency hopping (FH) technique. As is known it is difficult to design good FH code and synchronize the transmitter and receiver in a chaotic FH communication system. In recent years, application of chaos into the area of secure communication has attracted the research interest of many scholars. Having the statistical-random, noise-like, wide power spectrum distribution properties, easy to produce and sensitive to initial value, chaos application may improve the security of such communication systems. In this thesis, based on a simple piece-wise linear map, a chaotic FH code featuring good balance property, hamming correlation property and large FH span is developed. Computer simulation results show that this FH code is a good one when employed in a chaotic communication system.On the other hand, as Pecorra and Carrol demonstrated in 1990 that is able to synchronize chaotic systems, many chaos synchronization schemes have been proposed in the literatures. Our contribution in this regard is the following:A chaotic FH sequence is developed by using Logistic map as well as a multiple layer perceptron(MLP) neural networks;A fuzzy randomized sequence is designed and used in synchronizing the transmitter and receiver of a chaotic FH communication system.In Part Two of the thesis, a novel predistortion technique is addressed .As is well known that for wireless communication systems, to simultaneously enhance the power amplifier's efficiency and the utilization ratio of electro-magneticspectrum are two important but conflict each other aspects. One needs nonlinearity to improve the efficiency of power amplifier, on the other hand, the high spectrum efficient linear modulation method is very sensitive to power amplifier's nonlinearity. Such nonlinearity would result in a significant performance degradation of a communication system, such as constellation warping, BER deteriorating, intermodulation distortion and related unnecessary spectrum spreading, etc. So the amplifier needs to be linearized. In this thesis, different from the conventional predistortion technique, the power amplifier's predistortion is implemented by BP neural network, fuzzy neural network, and network parameter's adjustment by fuzzy training algorithm, respectively. At the same time, through neural network learning and error feedback between the input and actual output, this method is not only suitable for power amplifier's model but can also tolerate the amplifer's characteristic change which would results in temperature drift and component aging. In this sense, one can implement the adaptive power amplifier predistortion by using this technique .