基于PASTd的圆阵ESPRIT算法

An UCA-ESPRIT Algorithm Based on PASTd

基于旋转不变技术的信号参数估计(ESPRIT)算法因具有高效的波达方向(DOA)估计性能而受到广泛的研究和应用,但其需要进行协方差矩阵特征值分解,计算量大,不利于某些场合的具体工程实现。为此,提出将递推求解信号子空间的紧缩近似投影子空间跟踪(PASTd)算法与ESPRIT算法相结合,利用PASTd算法求出信号子空间特征向量,避免了特征值分解,降低了计算量与复杂度,并将其在圆阵中加以应用。所提算法不仅更适合硬件实现,而且将ESPRIT算法的应用进行了扩展。仿真结果验证了所提方法的有效性。

Due to its high-resolution and robust performance, ESPRIT algorithm has become the interest of many researches and applications in recent years. ESPRIT algorithm requires eigendecomposition of covariance matrix which has great computational complexity. It is inconvenient when the algorithm is used in specific engineering applications. This paper uses the PASTd algorithm to obtain the signal subspace iteratively and combine it with the ESPRIT algorithm, such that the eigendecomposition is avoided and the computational load and system complexity is reduced. The application of this method to the circular array is also discussed. The proposed algorithm is not only more suitable for hardware implementation, but also extends the application of the ESPRIT algorithm. Simulation results show the effectiveness of the method .