摘要
提出当DFT点数不是2的整数次幂时,用AFT计算SSCA算法中的DFT得到信号循环谱,并对AFT中的乘法进行改进.借助实数乘法和实数加法运算量的换算系数,用加法次数作为衡量算法计算量的标准,给出改进的AFT和DFT、FFT的计算量的闭合表达式,并用它们分析SSCA算法的计算量.最后,利用AFT对SSCA算法进行实现和性能仿真.仿真和试验结果表明:根据不同的分辨率,采用FFT或AFT计算SSCA算法中的DFT,能以最小的运算量实现任意分辨率的SSCA算法.
DFT contained by strip spectral correlation algorithm (SSCA) was computed by AFT to obtain the cyclic spectral when the length of DFT was not the integral power of 2.And multiplication of AFT algorithm was modified. By virtue of conversion parameter between real multiplication and real addition, the paper presented the closed-form expression of modified AFT, FFT and DFT employing the number of addition, and the computational complexity of SSCA was analyzed in terms of them. And SSCA was realized by AFT, through which its performance was simulated as well. Simulation results prove that SSCA with arbitrary resolution can be realized with minimum computational complexity if DFT is calculated using FFT or AFT according to different resolution.
出处
《哈尔滨工业大学学报》
EI
CAS
CSCD
北大核心
2008年第7期1040-1043,共4页
Journal of Harbin Institute of Technology
基金
国家高技术研究发展计划资助项目(2004AA001210)
关键词
AFT
循环谱
SSCA算法
计算量
arithmetic Fourier transform(AFT)
cyclic spectral correlation
strip spectral correlation algorithm
computation complexity