Design of Matrix Inversion Algorithm Based on Field Programmable Gate Array
Because the adaptive anti-jamming algorithm has a time delay when updating the optimal weight,it is difficult to meet the weight update rate requirements in a dynamic environment.In view of this situation,scholars have studied how to implement the fast sampling matrix inversion algorithm,but there is still a problem that it is only applicable to low-dimensional matrices and the weight update rate is slow.In order to solve the above problems,a sampling matrix inversion algorithm implementation architecture based on Cholesky decomposition was proposed.The implementation architecture mainly includes a covariance matrix calculation module,a Cholesky decomposition module,an inverse matrix module for calculating the lower triangular matrix L,a triangular matrix multiplication and weight calculation module.This design adopts a pipeline and state machine implementation structure,which effectively solves the problem of slow weight update rate caused by the large inverse operation of the high-order sampling matrix.The simulation results show that on the hardware platform of field programmable gate array(FPGA),for the 56-order sampling matrix,the update time of a weight only needs 1.2 ms under the working frequency of 100 MHz.The implementation architecture proposed in this article provides a practical and feasible solution for adaptive anti-jamming to quickly solve weight values,and has certain reference value for weight solving systems with similar requirements.
万方数据
维普
