Parameter estimation of a moving acoustic source: Linear chirplet transform vs WVD
Sai Gunaranjan Pelluri, T.V. Sreenivas · 2017
Doppler effect manifests in the form of changing instantaneous frequency (IF) of the received signal of a moving acoustic source. Several parameters characterizing a moving source are inherently embedded in the received IF. Thus, a careful analysis of the IF is required to obtain the source motion parameters. In this paper, we show how the IF captures the source information, such as speed, closest distance of approach, transmitted frequency and cross-over time, when a source moves along a linear path with a constant velocity. In such a scenario, the received IF is a non-linear function of time. Conventional approaches involve estimating the non-linear IF and then obtaining the source parameters by solving a non-linear least squares problem which is computationally very expensive. However, in this work, we show that the IF can be modeled by a linear function around the cross-over instant. We then estimate the source parameters using the slope and axis intercepts of the linear IF. In this regard, we use two time-varying signal analysis approaches to estimate IF namely the linear chirplet transform (LCT) and the Wigner-Ville distribution (WVD). We compare the performance of these two techniques in estimating the source parameters and show that the LCT is a better tool to estimate linear frequency modulations than the WVD.