徐项项, 吴阳, 苏杭, 刘金龙, 杨慧珍, 龚成龙. 基于FPGA的自适应光学系统SPGD算法实现[J]. 应用光学, 2021, 42(5): 810-816. DOI: 10.5768/JAO202142.0501008
引用本文: 徐项项, 吴阳, 苏杭, 刘金龙, 杨慧珍, 龚成龙. 基于FPGA的自适应光学系统SPGD算法实现[J]. 应用光学, 2021, 42(5): 810-816. DOI: 10.5768/JAO202142.0501008
XU Xiangxiang, WU Yang, SU Hang, LIU Jinlong, YANG Huizhen, GONG Chenglong. FPGA -based SPGD algorithm implementation of adaptive optical system[J]. Journal of Applied Optics, 2021, 42(5): 810-816. DOI: 10.5768/JAO202142.0501008
Citation: XU Xiangxiang, WU Yang, SU Hang, LIU Jinlong, YANG Huizhen, GONG Chenglong. FPGA -based SPGD algorithm implementation of adaptive optical system[J]. Journal of Applied Optics, 2021, 42(5): 810-816. DOI: 10.5768/JAO202142.0501008

基于FPGA的自适应光学系统SPGD算法实现

FPGA -based SPGD algorithm implementation of adaptive optical system

  • 摘要: FPGA(field programmable gate array)具有可重复编程、算法实现灵活等优势,以97单元变形镜为原型,提出基于FPGA的SPGD(stochastic parallel gradient descent algorithm)控制算法通用快速实现方案。采用TimeGen软件对SPGD算法进行时序分析,并采用Vivado软件分别对SPGD算法随机扰动电压生成、性能指标计算和控制电压的计算与输出进行FPGA配置和编程,将每个模块的计算结果与Matlab结果进行对比、分析。研究结果证明了该文所提方案的合理性和可行性,为下一步基于FPGA的SPGD算法硬件实现和应用提供基础。

     

    Abstract: The field programmable gate array (FPGA) has the advantages of repeatable programming and flexible algorithm implementation. With 97-element deformable mirror as the prototype, the general and rapid implementation scheme of the FPGA-based stochastic parallel gradient descent (SPGD) control algorithm was proposed. Firstly, the TimeGen software was adopted to analyze the time sequence of the SPGD algorithm. Secondly, the Vivado software was adopted to configure and programme the FPGA of the SPGD algorithm in the field of random disturbance voltage generation, performance indexes calculation, as well as control voltage calculation and output. Finally, the calculation results of each module and those of Matlab were compared and analyzed. The results prove the rationality and feasibility of the proposed scheme and provide the basis for the next step of hardware implementation and application of FPGA-based SPGD algorithm.

     

/

返回文章
返回