IMPLEMENTATION OF NOISE CANCELLATION WITH HARDWARE DESCRIPTION LANGUAGE

The objective of this project is to implement noise cancellation technique on an FPGA
using Hardware Description Language. The performance of several adaptive algorithms is
compared to determine the desirable algorithm used for adaptive noise cancellation
system. The project will focus on the implementation of adaptive filter with least-meansquares
(LMS) algorithm or normalized least-mean-squares (NLMS) algorithm to cancel
acoustic noises. This noise consists of extraneous or unwanted waveforms that can
interfere with communication. Due to the simplicity and effectiveness of adaptive noise
cancellation technique, it is used to remove the noise component from the desired signal.
The project is divided into four main parts: research, Matlab simulation, ModelSim
simulation and hardware implementation. The project starts with research on several noise
cancellation techniques, and then with Matlab code, Simulink and FDA tool, the adaptive
noise cancellation system is designed with the implementation of the LMS algorithm,
NLMS algorithm and recursive-least-square algorithm to remove the interference noise.
By using the Matlab code and Simulink, the noise that interfered with a sinusoidal signal
and a record of music can be removed. The original signal in turns can be retrieved from
the noise corrupted signal by changing the coefficient of the filter. Since filter is the
important component in adaptive filtering process, the filter is designed first before adding
adaptive algorithm. A Finite Impulse Response (FIR) filter is designed and the desired
result of functional simulation and timing simulation is obtained through ModelSim and
Integrated Software Environment (ISE) software and FPGA implementation. Finally the
adaptive algorithm is added to the filter, and implemented in the FPGA. The noise is
greatly reduced in Matlab simulation, functional simulation and timing simulation. Hence
the results of this project show that noise cancellation with adaptive filter is feasible.