The Study On FPGA Implementation And Applications Of Adaptive Filter

Adaptive filter with a stable filtering performance and good adaptability in an unknown environment gradually become a utility in areas such as digital signal processing and automatic control applications. At present, the research of adaptive filter focuses on two aspects:One aspect is in theory how to improve an adaptive algorithm to gain faster converge speed, smaller steady-state offset, lower computational complexity, better Robustness, etc.; the other aspect is how to use the hardware circuit to implement high performance filters and get practical applications in a variety of high-tech field. Therefore, the main work and contributions in the paper are as follows:(1) The principle of adaptive filtering, LMS algorithm and its various improved algorithms were studied, and an improved variable step size NLMS algorithm was proposed in this thesis. Through compared the Simulation results with LMS algorithm, NLMS algorithm, variable step size NLMS algorithm,it was found that the improved algorithm has a faster convergence rate and smaller steady-state offset. So, this illustrate that the improvement measure has improved the algorithm Overall performance.(2) The 8th-order adaptive variable step size NLMS filter was designed by Verilog HDL in this thesis. Some design techniques such as top-down method, module reuse, state machine were used in the design. Ultimately, Modelsim combined with Matlab was used to simulate the design. The results show the correctness of the design and the feasibility of adaptive filter FPGA implementation.(3)Based on the analysis of the hazards of harmonics in the power system and currently harmonic current detection methods, the improved algorithm was applied to detect harmonic.The simulation results show that when the harmonic or voltage frequency changed, the adaptive filter is still able to quickly detect the Fundamental wave and harmonic current from the non-linear load current, and the Harmonic distortion rate of the fundamental wave is less than 10%of the original. These illustrate that the adaptive harmonic detection method can detect the harmonics precisely and the algorithm is practical.