NLMS Adaptive Algorithm Implement Based on FPGA

Abstract

A adaptive filter was implemented in this paper, which was based on NLMS algorithm. High performance signals could be gotten by filtering the time-varying and unknown interference in the communication channels. The NLMS algorithm was achieved by discussing the principle of LMS algorithm and its improvements. It was concluded that NLMS algorithm could be implemented on FPGA chips. This paper described the method of the specific implementation. This method introduces bit-shift in terms of subsection instead of division operation, by which the operation speed of FPGA is improved apparently. The spectrogram of the output signals proved that the attenuation reached to 99.21 dB when the normalized frequency was 0.04375π offset from the center frequency. The adaptive filter could filter the interference effectively in the communication channels. Moreover, the implementation of this adaptive filter requires considerably less FPGA resources because of the decreases of its calculating complexity. It could meet the requirements of high-speed signal processing.