Study of TFR Techniques for Chirp Parameter Estimation

Chirp parameter estimation is the topic of research interest for various years. Nowadays, chirp signals are applied in many engineering fields such as radar and passive/active sonar systems. Related to the use of chirp signals, the problem of parameter estimation of the signal has been receiving a great deal of interest. Chirps signal is a non-stationary signal whose frequency varies with the time. Ordinary Fourier transform is inefficient due to strong non-stationary characteristics of the signal. Hence special techniques are required to evaluate a chirp signal. Literature survey shows that the parameter estimation techniques are broadly categorized as Time-Frequency representation and statistical representation. In this paper we will discuss some of the Time-Frequency representation techniques like FrFT, WVD and Bowtie in detail and their respective results are further evaluated. Simulations of some of the techniques are done with respect to various parameters taken into account. Comparative study based on the simulation results is also done. All these techniques have trade-off between the accuracy and complexity. The evaluation results show FrFT has more accurate results than WVD and Bowtie in presence of noise. That is noise immunity of FrFT is much more than WVD and Bowtie. But complexity is less in WVD and Bowtie is simplest to implement.