【摘要】 提出了一种改进的经验模式分解(EMD)分解方法,并将EMD方法应用于脉冲雷达速度与加速度估计。为了减小噪声对EMD分解和提取信号瞬时频率的影响,对分解后的信号进行区间阈值化处理并重构信号,对重构后的信号进行重新分解。对经改进的EMD分解得到的IMF作Hilbert变换可得到一组瞬时频率,对提取的瞬时频率进行多项式拟合得到的多项式系数分别对应于相位各阶系数,从而估计目标的径向加速度和加加速度。仿真结果表明,该方法在低信噪比时仍能得到很好的信号参数估计效果,在0 d B<SNR<15 d B时估计性能优于传统EMD方法,参数估计误差逼近C-R下界。最后通过脉冲雷达实测数据验证了文中方法的有效性。更多还原

【Abstract】 A modified Empirical mode decomposition( EMD) method is proposed,which is applied in pulse radar velocity and acceleration estimation. To alleviate the effect of noise on EMD and the extraction of instantaneous frequency( IF),the decomposed signal is processed with interval threshold to reconstruct the signal,and the reconstructed signal is decomposed again using EMD. A set of IFs is got through Hilbert transform on the newly decomposed IMF,which can be used to extract the frequency principal component of the signal. The frequency principal components are fitted with polynomials to get the coefficients of the phase,which can be used to estimate the radial acceleration and jerk. Simulations prove that the estimation performance of this method is well even under low SNR circumstances. The errors of estimated parameters are lower than traditional EMD when 0 d B<SNR<15 d B,and the estimation error is much closer to C-R lower bound of each parameter. The performances on measured pulse radar data prove the effectiveness of this method.更多还原