摘要
在空时二维自适应处理系统中对海量数据作脉压处理耗时颇大。本文基于单片数字信号处理器(DSP)从数据级和指令级两个方面进行优化,通过合理调度处理器资源并利用片内并行技术,快速实现了对海量数据的脉冲压缩优化处理。以空时二维自适应处理(STAP)系统中单景四重频数据为分析对象,比较了采用商业库优化方法和本文方法优化后的两种情况。实验结果证明,在处理时间上取得了1.801的加速比,误差控制在10-5量级,1 024点脉压处理时间为50.456μs,验证了本文方法的有效性。
Pulse compression processing time is serious for tremendous data in space-time adaptive processing （STAP） system. A fast implementation algorithm of tremendous data is presented by optimizing scheme from data level and instruction level based on single DSP, and by suitably scheduling the processor resource and using parallel techniques. The method for commercial library and the method of this paper are compared. The result shows that the speedup ratio is 1. 801, the error is controlled at the 10^-5 level, and the time of 1 024 points pulse compression is 50. 456 μs, thus verifing the validity of the method.
出处
《数据采集与处理》
CSCD
北大核心
2008年第3期342-346,共5页
Journal of Data Acquisition and Processing
基金
国家自然科学基金(60736009)资助项目
总装“十一五”预研(513070503)资助项目
