Cross-term Suppression Of The Wigner Distribution In Fractional Fourier Domain

In nature and engineering practice, we often encounter a number of time-varyingnon-stationary signals. Such as voice, biomedical signals, radar and sonar signals, andso on. Relying solely on the time or frequency domain analysis have been less thanthe requirements of engineering practice, The development of time-frequency analysistheory and techniques has provided an effective means of analysis and approaches tonon-stationary signals.Time-frequency analysis method can generally be divided into linear andnonlinear time-frequency representation. To time-frequency analysis of non-stationarysignals, although it does not produce cross-terms, but the time-frequency resolution islow, can’t satisfy the requirements of engineering practice. On the contrary, Thenon-linear time-frequency representation has a very high resolution, together withgood time-frequency properties, but trapped in cross-term interference by themulti-component non-stationary signals generated. For example: Wigner-Villedistribution (WVD). First, in this paper, to the cross-terms of multi-componentnon-stationary signals has been studied in time-frequency analysis. In-depth detailedanalysis of adaptive time-frequency distribution based on adaptive decomposition ofthe signal, with the close relationship of adaptive time-frequency distribution and theWigner distribution of the original signal, It get a improved time-frequencydistribution by the associated operations using the new adaptive time-frequencydistribution, combined with the Wigner distribution. Finally, on the basis oftheoretical analysis, the paper gives the algorithm implementation and analysis ofalgorithm performance. Simulation results show that the improved time-frequencydistribution is not only no window effect, with the same resolution and WVDdistribution, but also able to completely eliminate the cross-term interference.