Projects
ABSTRACT
In this project I have worked on FPGA-Implementation of adaptive filtering by using the concept of LMS Adaptive Filtering. First of all to do this a simple model was presented to show the general concept of working model of LMS Adaptive Filter. This model comes from the previous project that I have done titled “Simulation of LMS Adaptive Filters in noise cancellation”. In background section a brief introduction of this technique will be discussed so the following section could be more understandable. Later on a new model of LMS Adaptive filter will be discussed and be examined by running simulations and codes for testing the accuracy of new model. In new model I used the functional model of LMS Adaptive filter in which we have access to what is going inside LMS adaptive filter. It is good to mention that these simulations have been run using Simulink program, also to give a better sense of the difference between LMS algorithms, a sound signal used as a reference. Then, by adding a random noise generator, a noisy signal is generated. Finally by passing the noisy signal through the different LMS algorithms, we have the filtered signal. All different signals are audible ones. So not only by using the figures, but also by our hearing ability, we can compare the different algorithms. In addition to a numerical comparison of the simulation, a few programs have been written to illustrate the difference in numbers. After verification of functionality of the new model, in the following chapter I go through the implementation of the LMS functional model. A project file has been created in Xilinx ISE software that includes all the VHDL modules related to the functional model. Later on I’m going through synthesize and place and root. Finally by generating the programing file I have download the program to the FPGA, All of these steps have been shown by figures.
ABSTRACT
In this project I have done the simulation on adaptive filtering by focusing on the LMS Adaptive Filtering algorithm, and by setting up a couple of simulations and comparing the results and the usage of such algorithms in practice. To do this, first a very simple model was presented to give a general understanding of what is going on in this project, and every element was introduced to make it simple for everyone with a different level of knowledge about DSP. In order to have an idea of the results from the simulation, a small introduction of FIR filters is presented and then the general mechanism of the LMS adaptive filters is discussed. Later on, different features of the LMS filters are represented by showing the simulation results and by comparing LMS and Normalized LMS Adaptive filtering algorithms, which are the most used algorithms in real world situations. For this purpose the filter weights and the frequency response of the FIR filter have been chosen for comparison. In final simulation, different variations of the LMS adaptive filters are presented and simulated at the same time. Simulation results are illustrated to give a visual understanding of the pros and cons of each method. It is good to mention that these simulations have been run using Simulink program, also to give a better sense of the difference between LMS algorithms, a sound signal used as a reference. Then, by adding a random noise generator, a noisy signal is generated. Finally by passing the noisy signal through the different LMS algorithms, we have the filtered signal. All different signals are audible ones. So not only by using the figures, but also by our hearing ability, we can compare the different algorithms. In addition to a numerical comparison of the simulation, a few programs have been written to illustrate the difference in numbers.
More Academic Projects ...
