In search for the best wavelet for denoising low SNR RF Signal for FMCW Radar Altimeter

Kosgi, V (2014) In search for the best wavelet for denoising low SNR RF Signal for FMCW Radar Altimeter. Masters thesis, Indian Institute of Technology, Hyderabad.

Preview

Text
EE12M1040.pdf - Submitted Version
Download (3MB)

Abstract

This research project / thesis proposes the best wavelet for Denoising under low Signal to Noise Ratio (SNR) conditions and Discrete Wavelet Transform Architecture based design for RF signal denoising, targeting the real time applications like FMCW (Frequency Modulated Continuous Wave) Radar Altimeter used in Anti-Radiation Missiles, Smart Bombs, Fighter aircrafts, Helicopters etc., and other Defense and RF carrier based applications like Cellular communication. DWT (Discrete Wavelet Transform) & IDWT(Inverse DWT) Architecture Models designed in MATLAB for the wavelets under study like dmey, coif1, sym2, & debouches db1, db2, db3, db4, db6. The reconstructed signal results after denoising are compared in various aspects. The results show that db3 is the best wavelet for denoising application point of view. Finally the db3 based architecture design implemented in VHDL(VHSIC Hardware Description Language) and the simulation results compared, synthesis has been done using Xilinx ISE Design suite targeting an FPGA. This project involves study and implementation of De-noising algorithms using Discrete Wavelet Transform. In a system the signal to noise ratio (SNR) is important for reliable information retrieval. Analysis of signals with poor SNR may lead to wrong interpretation of results. Conventional techniques like filtering in time domain and frequency domain has its own limitations in estimating and characterizing noise. Wavelet transforms is a very useful tool in the analysis of non-stationary signals. Wavelet transform has been used in signal processing fields such as de noising or data compression. This method consists of decomposing the data recursively into a sum of details and approximations at different levels of resolution. The details represent the high frequency components while the approximations represent the low frequency components of the signal.

[error in script]

IITH Creators: