Analysis of Chirp and Multi-component Signals Using Signal Processing Methods

Abstract

Signals are represented by function in time or space. These are classified as stationary and non stationary signals. Stationary signals are those signals in which frequency, amplitude and phase do not change with time, whereas these spectral components change with time in non stationary signals. Chirp and multi-component signals are the examples of non stationary signal. Chirp signals contain frequency which either increases or decreases with time. On the hand, multi-component signals have more than one frequency component. The importance of extracting information out of chirp or multi-component signals lies in the fact that in several research fields, for example speech recognition, medical fields, radar signals, micro seismic signals, micro doppler signal detection, instantaneous frequency estimation etc., the signals are often multi-component, chirp or both. So to extract the information at a particular time instant corresponding to a particular frequency, therefore various time frequency tools have been developed for the analysis of chirp and multi-component signals such as Short time fourier transform (STFT), Wavelet transform (WT), Chirplet Transform (CT), Polynomial Chirplet Transform (PCT) and S-transform (ST). As it is known that STFT has a limitation of fixed window length, therefore to overcome this problem of STFT, WT has been reported. WT does not have the direct relationship with frequency. Therefore to provide direct relationship with frequency, S transform has been introduced and it also provides phase information. The S-transform can give information about the phase of each frequency, but it degrades the time resolution at lower frequencies and degrades frequency resolution at higher frequencies. So, modified S-transform has been proposed to improvise the performance of S-transform. In this dissertation, the performance of the proposed method has been compared with STFT, ST, CT and PCT for multi-component signals. The implementation of these signal processing tools have been done using MATLAB software.