基于FPGA实现快速矩阵求逆算法

Implementation of Fast Matrix Inversion Algorithm based on FPGA

Cholesky分解是一种矩阵运算方法。相比传统的矩阵求逆算法,它能够大大简化矩阵求逆的运算量,提高实时性。因此,介绍Cholesky分解原理及方法,并根据这一特性,在FPGA中实现基于Cholesky分解的快速矩阵求逆算法。FPGA具有流水线设计的特点,能够进一步提高接收抗干扰处理的实时性。用Matlab对FPGA实现的各种矩阵规模数据进行仿真,根据仿真结果和FPGA实际资源选取最优的FPGA实现方案。

Cholesky decomposition is a matrix operation method.Compared with the traditional matrix inversion algorithm,it can greatly simplify the calculation of matrix inversion and improve the real-time performance.Therefore,the principle and method of Cholesky decomposition are introduced,and based on this feature,a fast matrix inversion algorithm based on Cholesky decomposition is implemented in FPGA.FPGA has the characteristics of pipeline design,which can further improve the real-time performance of receiving anti-jamming processing.Matlab is used to simulate various matrix-scale data implemented by FPGA,and the optimal FPGA implementation solution is selected according to the simulation results and actual FPGA resources.