一种改进的四阶累积量测向算法

Improved Cumulant Direction Finding Algorithm with Fourth Order

针对降维的四阶累积量矩阵因采样快拍数有限而存在估计误差的问题,提出一种结合To-eplitz近似的四阶累积量旋转不变子空间(ESPRIT)测向算法(TFOC-ESPRIT).该算法在保持虚拟阵列有效孔径不变的同时,通过去除原始四阶累积量矩阵中的冗余元素,得到降维的四阶累积量矩阵,再对降维矩阵进行Toeplitz操作,恢复其Toeplitz结构,最后利用ESPRIT算法估计到达信号的方向.仿真实验表明:与MFOC-ESPRIT算法相比,TFOC-ESPRIT算法有效地提高了测向精度,不仅在少阵元数下能实现多信号的估计,而且当快拍数为400时,在空间白噪声和空间色噪声环境中其归一化成功概率也较高,分别达到了0.996和0.788.

An improved direction of arrival（DOA） estimation algorithm,called Toeplitz fourth-order cumulant ESPRIT（TFOC-ESPRIT） algorithm,is proposed to deal with the estimation error problem that exists in the reduced-rank fourth-order cumulant（FOC） matrix and is caused by finite sampling snapshots.A reduced-rank FOC matrix is obtained through removing the redundant information encompassed in the primary FOC matrix,while the effective aperture of the virtual array keeps unchanged.Then,the Toeplitz structure of the reduced-rank FOC matrix is recovered by applying the Toepltiz approximation.Finally,the direction of arrival signals is estimated using the ESPRIT algorithm.Simulation results and comparisons with the MFOC-ESPRIT algorithm show that the proposed TFOC-ESPRIT algorithm effectively improves direction-finding accuracy,and that it can not only estimate DOA of more targets with less antennas,but also have higher normalized probability of success-the probability reaches 0.996 and 0.788,respectively,in both the spatially-white noise and the spatially-color noise environments with 400 sampling snapshots.