Search

This paper addresses several issues associated with distribution and interpolation of time series, including model selection and various data transformations. Monte Carlo experiments are performed, which suggest that failure to account for these data transformations may lead to serious errors in estimation.

The paper addresses the problem of the estimation of timevariant Do pler centroid frequencies from SAR data. In particular tfe X-SAR sensor on the space shuttle is investigated. It is proposed to improve Doppler centroid estimates obtained from small azimuth segments by a Kalman smoother in along track direction. Ashuttle attitude model is developed to generate ical Doppler centroid histories. These fore they enter the Kalman smoother under test. Simulated Doppler estimates are shown to ether with the results obtained by Kalman smoothers wita different tuning parameters. histories are distur ?? ed by the assumed estimation errors be

Joint estimation of channel impulse response and its dynamic parameters using the expectation maximization (EM) algorithm and its MAP variant is derived for broadband shallow‐water acoustic communication channels. Based on state‐space channel modeling, the EM algorithms estimate the channel dynamic parameters from the sequence of channel impulse response estimates. The estimated parameters are then used in the Kalman smoother, which estimates the channel impulse response. The stability of the algorithm is shown to be related to an extended persistent excitation (EPE) condition, which requires that both the symbol sequence and the channel estimates be persistently exciting. Modified algorithms are proposed for broadband multipath channels to avoid the issue of insufficient excitation. Efficient suboptimal algorithms are also derived from the EM algorithms that alternatively estimate the parameter and the channel impulse response while allowing slow parameter variations. The performance of these channel est...

This paper presents a fully Bayesian hierarchical model for blind audio source separation in a noisy environment. Our probabilistic approach is based on Gaussian priors for the speech signals, Gamma hyperpriors for the speech precisions and a Gamma prior for the noise precision. The time-varying acoustic channels are modelled with a linear-Gaussian state-space model. The inference is carried out using a variational Expectation-Maximization (VEM) algorithm, leading to a variant of the multi-speaker multichannel Wiener filter (MCWF) to separate and enhance the audio sources, and a Kalman smoother to infer the acoustic channels. The VEM speech estimator can be decomposed into two stages: A multi-speaker linearly constrained minimum variance (LCMV) beamformer followed by a variational multi-speaker postfilter. The proposed algorithm is evaluated in a static scenario using recorded room impulse responses (RIRs) with two reverberation levels, showing superior performance compared to competing methods.

A time-varying parametric spectrum estimation method for analyzing nonstationary heart rate variability signals is presented. In the method, the nonstationary signal is first modeled with time-varying autoregressive model and the model parameters are estimated recursively with a Kalman smoother algorithm. The spectrum estimates for each time are then obtained from the estimated model parameters. Statistics of the obtained spectrum estimates are derived using the error propagation principle. The obtained spectrum estimates can further be decomposed into separate components and, thus, the time-variation of low and high frequency components of heart rate variability can be examined separately.