Design and implementation of least mean square adaptive filter on Altera Cyclone II Field Programmable Gate Array for active noise control

Abstract

This paper presents a design and implementation of least mean square (LMS) adaptive filter for use in active noise control (ANC) application. The filter has been designed and synthesized with Altera Quartus II development platform and implemented on Cyclone II FPGA available on Altera DE2 development board. Architecture of the adaptive filter is based on conventional multiplier-adder to realize the multiply and accumulate (MAC) operation for FIR, implementing 64 filter tap with 9- bit signed integer coefficients. A uniform random signal with the bandwidth of 24 KHz is used as the test and reference signal for the filter verification. This signal is generated internally on FPGA based on LFSR technique. Performance of the filter is evaluated by means of convergence speed (with different step size, µ ) and comparison of adaptive filter coefficients with the unknown system. A fixed filter is programmed to behave as a bandpass FIR Hamming-type window with the cut-off frequency from 250Hz to 3400Hz to act as the unknown system. The results conclude that the LMS adaptive filter designed have successfully adapted with the unknown system with the fastest convergence speed of 1.46 ms at 1/2 <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">10</sup> step size.