中文摘要：

基于局部多项式傅里叶变换（local polynomial Fourier transform，LPFT），提出了一种新颖有效的窄带多项式相位信号波达方向估计算法。虽然LPFT比Wigner-Ville分布运算复杂，但该变换具有更优的瞬时频率估计能力、较低的交叉项影响，及处理多项式相位信号的能力。理论分析表明，本文方法的信噪比提升能力优于基于Wigner—Ville分布的传统时频子空间多重信号分类（multiple signal classification，MUSIC）法。仿真实验验证了该方法的性能。

英文摘要：

Based on local polynomial Fourier transform （LPFT）, a novel efficient narrow-band polynomial phase signal direction of arrival estimation algorithm is presented. Although LPFT is more complex than the Wigner-Ville distribution, it estimates the instantaneous frequency （IF） of signals more accurately, with lower cross-term effects and the ability to deal with polynomial phase signals. Theoretical analysis shows that the signal to noise improved ratio of the method is better than traditional time-frequency multiple signal classification （MUSIC） method based on Wigner-Ville distribution. Simulation results verify the performance of the method.