الفهرس 
introductionOnly 14 pages are availabe for public view
 |  | from | 175 |  |  |
| | | from | 175 | | |
Abstract
Noise cancellation is a very important topic nowadays. It has various usages in many applications such as communication, industrial and environmental protection, military application and many others.
Our purpose in this thesis is to describe the concept of adaptive digital filter in nOise canceling, to provide a theoretical and a practical treatment in Adaptive Digital Filters and Neural Networks. Study their advantages and limitations. Make a comparison between two techniques and get a conclusion ..
The first technique is using the LMS algorithm to estimate the parameters of the filter based on these parameters estimated canceling signal is generated. Study the effects of changing parameters of the LMS algorithm on the performance of the filter. Implement an adaptive noise canceling (ANC) filter with the DSP 56001 processor using LMS algorithm and study the performance of the filter with noise level variation.
The other technique is using the neural networks in noise canceling and signal identifier with back propagation learning algorithm that can discover the parameters of the network automatically through supervised learning.
Neural Networks and digital computers differ in two fundamental modes in computation and design. In a neural network based system design there is no need to 0 formula to describe a problem. This feature allows neural networks to solve problems that are very difficult to express mathematically. Our studies show that the performance of the noise canceling using neural network with back-propagation algorithm is better than adaptive digital filter with least mean square algorithm. So neural network seems to replace digital computer system in many of applications including noise canceling.