SCIENCE CHINA Information Sciences, Volume 62 , Issue 2 : 029304(2019) https://doi.org/10.1007/s11432-018-9491-y

A multicomponent micro-Doppler signal decomposition and parameter estimation method for target recognition

More info
  • ReceivedMar 29, 2018
  • AcceptedJun 13, 2018
  • PublishedOct 22, 2018


There is no abstract available for this article.


[1] Bai X R, Xing M D, Zhou F. Imaging of micromotion targets with rotating parts based on empirical-mode decomposition. IEEE Trans Geosci Remote Sens, 2008, 46: 3514-3523 CrossRef ADS Google Scholar

[2] Kang W W, Zhang Y H, Dong X. Micro-Doppler effect removal for ISAR imaging based on bivariate variational mode decomposition. IET Radar Sonar Navig, 2018, 12: 74-81 CrossRef Google Scholar

[3] Vishwakarma S, Ram S S. Detection of multiple movers based on single channel source separation of their micro-Dopplers. IEEE Trans Aerosp Electron Syst, 2018, 54: 159-169 CrossRef ADS Google Scholar

[4] Stankovic L, Djurovic I, Thayaparan T. Separation of target rigid body and micro-doppler effects in ISAR imaging. IEEE Trans Aerosp Electron Syst, 2006, 42: 1496-1506 CrossRef ADS Google Scholar

[5] Yang Y, Peng Z K, Dong X J. General parameterized time-frequency transform. IEEE Trans Signal Process, 2014, 62: 2751-2764 CrossRef ADS Google Scholar

[6] Sun Z S, Wang J, Zhang Y T. Multiple walking human recognition based on radar micro-Doppler signatures. Sci China Inf Sci, 2015, 58: 122302 CrossRef Google Scholar

  • Figure 1

    (Color online) Results of (a) STFT, (b) component 1, (c) component 2, (d) component 3.


    Algorithm 1 Multicomponent kernel function estimation method


    Require:$\xi$, $i~=~0$, ${P_{\rho,0}}~=~\{~{{a_{\rho,0}},{b_{\rho,0}},{\omega~_{\rho,0}}}~\}~=~0$, $\Gamma(r_h,\omega_h,\theta_h)$ $=0(r_h\in[r_{\rm~hmin},r_{\rm~hmax}]$,

    $\omega_h\in[\omega_{\rm~hmin},\omega_{\rm~hmax}]$, $\theta_h\in[\theta_{\rm~hmin},\theta_{\rm~hmax}])$.

    Output:Kernel function parameters ${P_{\rho~,i}}$ and the micro-Doppler signal frequency curve.

    For $i$-th Step:

    Get ${\rm~PTF}(t,\omega~;{P_{\rho~,0}})$ by parameterized TF analysis;

    Do the Hough transform on the TF domain in step 1, and find the local maximum value point in $\Gamma~({r_h},{\omega~_h},{\theta~_h})$ as $\Gamma~({r_{h\rho~}},{\omega~_{h\rho~}},{\theta~_{h\rho~}})$, the number of local maxima is $M$;

    $\rho~=~\rho~+~1$, $i~=~1$, ${P_{\rho~,i}}~=~[~{{a_{\rho~,i}},{b_{\rho~,i}},{\omega~_{\rho~,i}}}~]~=~[~{r_{h\rho~}}\sin~{\theta~_{h\rho~}},$ $-~{r_{h\rho~}}\cos~{\theta~_{h\rho~}},{\omega~_{h\rho~}}~]$;

    Get the peak ridge ${\hat~f_{m\text{-}D,\rho~,i}}(t)$, and calculate $\phi~_{S,\rho~,i}^R(a,$ $b,\hat~\omega~;\tau~)$ and $\phi~_{S,\rho~,i}^T(a,b,\hat~\omega~;\tau~,t)$;

    Calculate ${P_{\rho~,i}}$, get ${\rm~PTF}(t,\omega~;{P_{\rho~,i}})$ by parameterized TF analysis;

    Calculate ${\Lambda~_{\rho~,i}}$ using (7);

    If ${\Lambda~_{\rho~,i}}~>~\xi~$, then $i~=~i~+~1$, ${P_{\rho~,i{\rm{~+~}}1}}~=~{P_{\rho~,i}}$, and go to step 4; else, and if $\rho~\le~{\rm~M}$, go to step 3.END