Performance Analysis of Time Frequency Resolution Techniques for Non-Stationary Signals

Background/Objectives: All Traditional data analysis methods are based on linear and stationary assumptions. One of the important attributes of the signal processing techniques are applicable for non stationary signal environments is an improvement in time-frequency resolution or localization, relative to classical techniques like Fourier Transform (FT) which overcome these drawbacks of FT. The main objective of this paper is to improve the target detection performance in active sonar and radar based systems. Methods/Statistical Analysis: In this paper a study of time-frequency resolution analysis of non stationary signals using different transform techniques like Short Time Fourier Transform (STFT), Continuous Wavelet Transform (CWT) and Fractional Fourier Transform (FrFT) is presented. Findings: STFT provides the time informationby computing multiple FFT’s for consecutive intervals of time and then putting them together. The spectrogram can able to resolve the temporal frequency evolution content from the signal. Spectrogram has a trade-off in time and frequency resolution in accordance with the uncertainty principle of Heisenberg. For the Short Time Fourier Transform, time-frequency resolution is fixed. It can be varied in the Continuous Wavelet Transform as a function of an analyzing frequency. The Continuous Wavelet Transform having the analysis function can be chosen with more freedom. Fractional Fourier Transform is a time-frequency distribution which provides us with an external degree of freedom. It can allow signal to be transformed into a fractional domain with a fractional order parameter α. Application/Improvement: Time Frequency resolution or localization transform techniques are used to improve the performance of target detection in active sonar and radar systems.