Research On Adaptive Blind Equalization Algorithm For Shortwave Communication Channels

Short-wave communications is widely used in communication field, especially in military communication, because of its simple equipments, facility-employment and the communication capability with long distances. However, because the multi-path effects cause severe inter symbol interference (ISI), the performance and capacity of communications is not very satisfactory, researchers and engineers do their best and propose many solutions to eliminate the ISI, among which the equalization technology is always the key and challenging task. In recent years, blind equalization technology became a very hot research topic in communication and signal processing fields; because it doesn't need to transmit the training sequences and it can increase the capacity of communication channel greatly. The blind equalization technology has much potential application in future short-wave communications and wireless communication systems, the adaptive blind equalization algorithms of short-wave communication channels will be researched in this thesis.In the first chapter, the importance of short-wave communications is emphasized in military field. The impact of short-wave propagation channels on the performance of communications is reviewed, based on which, the necessity of blind equalization can be concluded. Then the current situation of blind equalization is summarized, the existed problems and their solving ideas are pointed, from which the aim of the thesis can be found. In the following chapter, the basic characteristic of short-wave communication channel is implicitly introduced. Then some characteristic of impact on short-wave signal transmission imposed by ionosphere are analyzed, such as multi-path interference and fading, which make that the impact on signal transmission caused by short-wave communication channels can be known more widely and more deeply. In the third chapter, the blind equalization's basic ideas is simply presented and the criterion on how to evaluate the blind equalization performance is introduced, and the main factors of evaluating the blind equalization algorithm which needed to be considered are summarized, which make of the theory foundation of the following researching work in this thesis.In the fourth chapter, a new blind equalization algorithm for short-wave signal system is designed which combined the maximum kurtosis criterion of the blind equalization of the linear system with the decision directed least mean error in the cost function. There are two obvious advantages of the algorithm. First the higher-order cumulating is adopted which can effectively restrain to Gaussian noise. Second, the genetic algorithm of the real number code is used in order to conduct the superior search of the criterion function, and the local optimum values problem that the traditional gradient searching method will be easily caught in partial convergence is solved in the algorithm. The conclusion of the algorithm is proved by simulation results.In the fifth chapter, a variable step size blind equalization algorithm is designed, whose error-controlling function switch between modified constant modulus algorithm and decision directed-least mean square error. The algorithm makes full use of the merits of both algorithms, accelerates the convergence rate and reduces the stable errors after the algorithm converging. In addition, the phase rotation problem occurred in constant modulus algorithm is solved through employing the cost function which contains phase information; at the same time the amplitude is also equalized.The conclusions and review of this thesis are made in the last chapter, and some research topics along the lines of this thesis in the future and the possible solution idea are pointed out.