Study And Implementation Of Adaptive Channel Equalization Algorithm For Wireless Communication System

In wireless communications, wireless channel multipath propagation will form the multi-path effect that impacts on inter-symbol interference(ISI) of the transmitted signal. With the development of communications technology, wireless communication transmission rate is higher and higher, the more severe multipath propagation effects of inter-symbol interference affect reception quality. To reduce the ISI and to improve the quality of communication, adaptive equalization can be adopted to remove multipath propagation effects. Adaptive channel equalizer is normally set at the receiving end. The equalizer can compensate for lack of channel characteristics based on the received signal adaptive to reduce ISI caused by multipath propagation, thereby improving accuracy of data transmission. The core of adaptive equalizer is the adaptive algorithm. A good equalizer should have high-speed convergence, strong tracking performance, etc, therefore the study of adaptive algorithm and adaptive equalizer has important theoretical and practical significance.The main point of this thesis is to study the structure and performance of the adaptive equalizer algorithm. Firstly, based on the study of wireless communication channel characteristics and channel simulation model, the thesis analyses the generation of ISI, the origin of equalizer and the structure and classification of equalizer. Then this theses studied the least mean square(LMS) algorithm, decision feedback equalization(DFE), recursive least squares(RLS) algorithm and the constant modulus algorithm(CMA). The theses also compared the performance of each algorithm, the performance of the algorithms under different parameters and the performance of each algorithm under different channel environments through a large number of simulations. Next, the thesis concentrates on the convergence of the LMS algorithm and analyses the factors which impacts on the convergence of the LMS algorithm factors. This thesis presents three modified LMS algorithms against different factors, and the simulations between proposed algorithms in this thesis and some existing improved algorithms demonstrate the superiority of the proposed algorithms in this thesis. Finally, the variable step size LMS algorithm was implemented by pipeline method on FPGA.In addition, due to all parameters of the LMS algorithm are sensitive to quantization, the difficulty of digital implementation increases. The long iteration path and feedback loops of the LMS algorithm lead to the implementation by pipeline method also more difficult, and the length of the pipeline will also affect other parameters. In this regard, we use the relaxation transformation technology in advance to implement the variable step size LMS algorithm pipelined.