基于激光导引头信号的并行高速FFT算法设计

Design of parallel high-speed FFT algorithm based on laser seeker signal

为了减少激光半主动武器中测量光学器件光斑点坐标时噪声和干扰对探测精度影响、增加脉冲信号的测量带宽、提取信号的有效值,同时克服串行快速傅里叶变换(FFT)运算耗时及时间复杂度较大的问题,基于多核和并行架构的SoC-FPGA平台以及OpenCL软件,提出了实现并行FFT的计算方法。结果表明,利用该方法可使FFT(1-D)的时间复杂度下降到原来的1/Q,得到了较好的加速效果;通过3种平台(先进精简指令集微处理器、数字信号处理器和片上系统现场可编程门阵列)的运算耗时实验对比,该算法运算耗时为6.0449ms(1-D4096点),要比同点数其它两种平台运算耗时少。并行FFT算法不仅满足激光半主动导引头信号实时性的要求,而且可以达到去噪的效果,能有效地降低噪声和背景光的影响。

In order to reduce the influence of noise and interference on the detection accuracy of optical spot coordinates in semi-active laser weapons,the measurement bandwidth of pulse signal increased and the effective value of the signal was extracted.By overcoming the great time-consuming and complexity of the serial fast Fourier transform(FFT)operation,parallel FFT computing method was proposed based on multi-core and parallel architecture system on chip-field-programmable gate array(SoC-FPGA)platform and OpenCL software.By this method,the time complexity of FFT(1-D)can be reduced to 1/Q times and the better acceleration effect was obtained.After comparing the computational time-consuming experiments of three platforms(advanced risc machines,digital signal processing and SoC-FPGA),the caculating time of the proposed algorithm is to 6.0449ms(1-D 4096 points)and less than that of the other two platforms with the same number of points.The results show that parallel FFT algorithm not only meets the requirement of the real-time performance of laser semi-active seeker and achieves the effect of denoising,but also can effectively reduce the influence of noise and background light.