FPGA Implementation Of High-order Complex Notch Filter With Constrained Poles And Zeros And Its Application

During transmission in a complex communications environment,spread spectrum signal is susceptible to be interfered by strong signals, and most of which are usually considered to be narrowband interference.In face of strong narrowband interference or multiple narrowband interference, the spreading gain alone to suppress interference is not feasible. Studies indicated that the performance of the spread spectrum system can be improved a lot by suppressing interference before de-spreading,especially for strong single frequent interference or narrowband interference.However in response to different narrowband interference,the interference suppression performance of notch filter with conventional structure varies very widely. To solve the problems above, high-order notch filter throries were deeply studied in this theise, the main research results obtained are as follows:1.Notch filter and the related theories are studied. Through research and analysis of transform-domain filter theories,time-domain of FIR and IIR filter theories, the high order complex notch filter with constrained poles and zeros algorithm is presented.Through the analysis of the first order notch filter, a general structure of the complex high-order matrix adaptive notch filter is promoted with the first order notch filter as the basic unit, and the fundamentals and the iteration steps of the high-order notch filter are also involved. For this design, the n-order complex notch filter can deal with n narrowband interference of different intensities at different positions in different SNR perfectly, while the effect on useful signal is very little.2. The structure of the notch filter is optimized. According to the structural features of the notch filter combined with Matlab simulation verification, several key factors and its impact on the system of the notch filter has been analyzed in detail. The step size greatly affects the convergence speed and convergence stability;the value of the pole radius directly determines the bandwidth of complex adaptive notch filter, which affects the notch filter performance and the iteration the convergence rate as well;the accuracy of the convergence point decide the accuracy of the frequency tack point and the steady results. To solved these problems, the dynamic-step and fixed-step combination method, the sliding window method and the separate parameters method are raised. Simulation show that the convergence speed of the iterative algorithm is greatly improved, and the convergence stability also increased a lot.3.The FPGA(Filed Programmable Gate Array) design of the first-order, second-order, and third-order notch filter is completed. For Matlab quantization, in order to reuse the modules, namely when we realize the high-order notch filter,the modules of the first order notch filter module can be directly obtained, we quantified the program synchronously. Based on the quantified program,we give a reasonable module division of the notch filter system, and for each module we conduct the FPGA design and verification with Modelsim simulation one by one.4. Complete the first-order, second-order, third-order notch board-level debugging and verification of FPGA design.Two verification approches are presented.First we use Signaltap sampling method, by drawing on the board-level data of the final results and intermediate data for verification, and then another approach by logic analyzer direct observation is also promoted, Both of the methods confirmed the correctness and validity of the FPGA design.