Established in 1991. We have celebrated 20 years of the group publishing research monography covering more than 150 papers published in leading world journals (about half in the IEEE issues) that are cited about 2500 times. Group is established by Prof. Ljubiša Stanković, IEEE Fellow and Associted editor in several IEEE Transactions and Letters as well as member of IEEE Technical Committee for Theory and Methods in Signal Processing, leading or guest editor of several special isues etc. Senior researchers invloved in the Laboratory are: › Prof. dr Igor Đurović › Prof. dr Miloš Daković › Assist. Prof. dr Slobodan Đukanović › Assist. Prof. dr Vesna Popović Bugarin › d r Marko Simenuović
Prof. dr Ljubiša Stanković Published in 2011
Time-frequency signal analysis with developed numerous inovative tools. Signal analysis, detection and estimation. Robust signal processing for impulsive and high noise environments. Applications.
Professor Stanković has proposed technique coined as the S -method that combines favorable preperties of various TF representations (high resolution, removal of spurios components, low calculation burden, good properties for noise environments). This technique has been used for development of various other TF represenations and for different applications. This technique can be asumed to have two simple steps: › Evaluation of the short time (windowed) Fourier transform (STFT) › Additional block adding corelated STFT samples (with rather small burden).
Spectrogram Wigner distribution S-method
Several novel signal analysis tools are proposed particularly for nonstationary signals. Techniques for nonparametric signal estimation based on the intersection-of-the-confidence intervals (ICI) (for precise estimation in the case of moderate level of noise) and the Viterbi algorithm based estimator (for low SNR) are proposed. In the area of the parametric estimator we have considered various modification of the Cubic Phase Function (CPF) Estimators , and application of meta-heuristic search techniques for estimation of the signal parameters including genetic algorithms . We have also proposed technique called quasi ML that can combine properties of both parametric and nonparametric estimators.
Ilustration of the ICI algorithm Adaptive parametric estimate obtained from various alternatives by simple comparisons Adaptive parameter ICI algorithm is used for IF estimation, features extraction, variable step LMS algorithms, image processing etc.
We have proposed various solutions for spectral analysis of signals corrupted by impulsive noise. Various robust DFT forms and TF representations are proposed. Some of these transforms are proposed in adaptive form with selection of the design parameter of the algorithms. We have proposed techniques for precise estimation of signal parameters from robust spectral analysis tools.
The main direction for applications in the past couple of years was related to radar signal processing mainly through the string of projects realized for DoD Canada. › Micro-Doppler removal and analysis › Focusing of SAR/ISAR images › Human gait extraction and following from radar images › Ship monitoring › Imaging and decomposition of radar image with low flying aircrafts › OTHR systems Other applications from communications and OFDM systems, car engine signals analysis, underwater signals processing, to power systems etc.
TF signature of received signal at the OTHR radar Decomposed signal from the signature where three signal components clearly appears.
Original defocused image Focused image #1 Adaptive parameter Focused image #2
TFR of received signal (log) TFR of target signal Eigenvalues 1 0.8 0.6 frequency frequency 0.4 0.2 0 0 5 10 15 20 time time number Fourier transform target velocity measure 12 100 10 50 8 amplitude 0 6 4 -50 2 -100 0 -0.5 0 0.5 0 MTr 0 5 10 15 20 normalized frequency time number
1 2 3 4 5 6 7 8 9 10 11 12
This is rather brief overview of the research results achieved in our group in past 20 years with numerous other research results achieved. Group is active in various international projects. Group is active in teaching both on undergraduate and graduate levels. We have published several textbooks for students covering various topics. More details on research activities can be found on www.tfsa.ac.me.
Recommend
More recommend
