基于FPGA的自适应阵列协方差矩阵特征分解的实现

卫星导航抗干扰的过程中，对空间信号波达方向估计、干扰个数检测、最优权矢量的求解直接影响着导航接收机的抗干扰性能，而协方差矩阵的特征分解是这些算法实现的核心部分。根据自适应阵列天线获得的协方差矩阵的特性，基于双边并行Jacobi算法，实现了基于FPGA的协方差矩阵特征值和特征向量的求解，并通过在信号波达方向估计的应用进行了验证。另外，在实现的过程中对直接调用CORDIC IP核的方式进行了精度误差分析，并用一种双精度浮点的方式进行修正，提高了矩阵特征分解FPGA的实现精度，为导航抗干扰接收机性能的提升提供了有效的工程基础。

Realization of Adaptive Array Covariance Matrix Feature Separation Based on FPGA

In the process of satellite navigation anti-jamming, the anti-jamming performance of the navigation receiver is affected by the estimation of jamming direction, the detection of jamming number, and the solution of optimal weight vector. The eigendecomposition of covariance matrix is the core of these algorithms. The characteristics of the covariance matrix obtained from adaptive array antennas are discussed in this paper. The eigenvalue and eigenvectors of adaptive array antenna covariance matrix based on FPGA are solved by the parallel Jacobi algorithm. It is verified by the direction of arrival estimation. In addition, the accuracy error is analyzed on the way of directly calling the CORDIC IP core and corrected by a double precision floating point method. The accuracy of matrix factorization implemented by FPGA is improved. It provides an effective engineering foundation for improving the performance of the navigation anti-jamming receiver.