IF and GD estimation from evolutionary spectrum
Signal Processing - Special section on Markov Chain Monte Carlo (MCMC) methods for signal processing
Zero-Tracking Time-Frequency Distributions
ICASSP '97 Proceedings of the 1997 IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP '97)-Volume 3 - Volume 3
Adaptive window in the PWVD for the IF estimation of FM signals in additive Gaussian noise
ICASSP '99 Proceedings of the Acoustics, Speech, and Signal Processing, 1999. on 1999 IEEE International Conference - Volume 03
IEEE Transactions on Signal Processing
IEEE Transactions on Signal Processing
Hybrid linear/quadratic time-frequency attributes
IEEE Transactions on Signal Processing
Improved instantaneous frequency estimation using an adaptiveshort-time Fourier transform
IEEE Transactions on Signal Processing
Time--frequency feature representation using energy concentration: An overview of recent advances
Digital Signal Processing
Hi-index | 0.08 |
Nonparametric algorithm for the instantaneous frequency (IF) estimation, by using the Wigner distribution (WD) with an adaptive window length, is considered in the paper. This algorithm produces a bias-to-variance trade-off close to optimal, meaning almost minimal mean squared error (MSE) of the estimation. Thus, the adaptive window length is characterized by a small bias at the considered instant. Then, according to the WD concentration property, the IF estimate can be assumed as a linear function within this window. Instead of nonparametric IF estimation in other points within this interval the linear IF interpolation can be performed. Length of the interpolation segment is determined based on the adaptive window length. It is done in such a way to produce a trade-off between the interpolation caused error and calculational complexity. This modification can produce a significant calculation savings, without increasing the overall MSE. Theoretical analysis has been confirmed on numerical examples and statistical study with four synthetic signals. The approach presented here can be generalized in a straightforward manner to nonlinear interpolations and higher order time-frequency representations.