Interpolation and sparse reconstrunctions of Doppler and microDoppler signatures under missing samples

The paper considers Doppler and microDoppler radar signature estimations in the presence of missing samples using time-frequency distributions. Incomplete or random sampled data can be due to ranging and localization enhancements, discarding noisy measurements, hardware simplification, sampling rate limitations, or logistical restrictions on data collections and acquisition. We demonstrate that the use of interpolators to estimate the missing samples in the instantaneous autocorrelation function outperforms time-domain data interpolations. We compare time-frequency distributions with and without data interpolations and contrast their performance with sparse signal reconstruction which exploits the sparsity of the Doppler signature in the time-frequency domain.