Frequency estimation of low rate non-uniformly sampled signals using neural networks

This paper reports on the employment of low rate non-uniform sampling to increase the unambiguous bandwidth in the frequency estimation of narrowband signals. First, we explore the effect of non-uniform sampling on the spectra of the sampled signals. We then show how non-uniform sampling affects the spectral estimation, when conventional methods are used. Finally, we present an effective neural network scheme and report on some simulation results. The results show that the instantaneous frequency content of the narrow-band signals may be estimated effectively, using low rate non-uniform sampling and the proposed neural network frequency estimation method. One important application of the proposed method is in unambiguous Doppler estimation of target echoes in long range radar with staggered pulse repetition frequency