A Parallel Processing Technique for Clutter Suppression of Passive Radar Based on LDLT Decomposition
With the increasing bandwidth of the opportunity exposure source signal,improving the detection ability of passive radar also leads to an increasing computational complexity of clutter suppression.Therefore,real-time clutter suppression has become a serious challenge faced by passive radar.In response to this,an improved Extended Cancellation Algorithm Batches(ECA-B)is designed based on LDLT decomposition and inter-segment parallel algorithm.Firstly,based on the multithreading parallel processing technology of the Graphic Processing Unit(GPU)and the same submodule characteristics of each segment of the ECA-B,the efficiency of the ECA-B is improved through segmented parallel processing;Then,in response to the time-consuming problem of data transmission in the inversion process of the traditional ECA-B algorithm,a parallel iterative inversion method based on LDLT is proposed using the conjugate symmetry property of the autocorrelation matrix.The inversion process is achieved through two Compute Unified Device Architecture(CUDA)kernel functions,saving data transmission time in the matrix inversion process and further improving the implementation efficiency of inter-segment parallel algorithms.The experimental results show that the proposed algorithm has higher timeliness and effectiveness compared to traditional algorithms.
国家科技期刊平台
NSTL
万方数据
维普
