镱原子光钟基于FPGA控制系统的数据处理  

Data Processing Based on FPGA Control Unit in Ytterbium Atomic Clock

作　　者：蔡禹 王进起 殷妮 贺凌翔 CAI Yu;WANG Jin-qi;YIN Ni;HE Ling-xiang(State Key Laboratory of Magnetic Resonance and Atomic Molecular Physics,Innovation Academy for Precision Measurement Science and Technology,Chinese Academy of Sciences,Wuhan 430071,China;Key Laboratory of Atomic Frequency Standards,Innovation Academy for Precision Measurement Science and Technology,Chinese Academy of Sciences,Wuhan 430071,China;University of Chinese Academy of Sciences,Beijing 100049,China)

机构地区：[1]中国科学院精密测量科学与技术创新研究院,波谱与原子分子物理国家重点实验室,湖北武汉430071 [2]中国科学院原子频率标准重点实验室,湖北武汉430071 [3]中国科学院大学,北京100049

出　　处：《量子光学学报》2021年第3期207-212,共6页Journal of Quantum Optics

基　　金：国家自然科学基金(U20A2075,11803072)。

摘　　要：高精度光钟被认为是广义相对论验证,相对论测地学以及暗物质探测的重要工具.镱原子光钟是目前国际上稳定度最好的光钟,通过36 h的平均可以进入10^(-19)的量级.控制系统是镱原子光钟的重要组成部分,它主要负责光钟系统的时序控制.本文提出一种基于Python语言支持的现场可编程门阵列(FPGA)控制系统,通过编译和调用Python程序包在FPGA上运行,用来产生光钟运行过程所需的各种控制信号,以及对镱原子光钟工作过程中获取的数据进行存储和后期处理.研究结果表明:基于FPGA控制系统的数据处理可以完全匹配传统Matlab和Origin软件的数据处理方法,获取μK量级冷原子的温度和Hz量级的钟跃迁谱线宽;另外,对复杂钟跃迁谱线实现高斯基函数的多峰拟合,通过与早期实验过程中噪声分析建立的数据库进行比对,用来作为光钟运行的过程分析和故障诊断,有助于未来将机器学习引入高精度原子光钟的数据处理用来实现引力波以及暗物质的探测.High accuracy optical clock is regarded as an important tool for testing general relativity,relativistic geodesy and dark matter detection.So far ytterbium optical clock shows the best stability,which reaches 10-19level after averaging 36 h.Control unit is the vital part of the ytterbium optical clock.It mainly takes on time sequence generation for the clock system.In this paper,a field programmable gate array(FPGA)control unit based on Python language is introduced to ytterbium optical clock.On one hand,it can generate various control signal for ytterbium optical clock by compiling and calling the subroutine.On the other hand,it can be used for data storage and processing during the clock operation.The results show that data processing with the FPGA system can be perfectly matched with traditional method dealt with Matlab and Origin program,and cold atoms with micro Kelvin level and clock transition spectra with Hertz level are obtained in ytterbium clock system.More importantly,based on the noise analysis database in earlier experiment,fitting to the complex clock transition spectra with multi-Gaussian superimposition function can act as a method for machine learning,which is a better way for experimental process analysis and fault diagnose.It would benefit for gravitational wave detection and hunting for dark matter with more accurate atomic optical clock based on machine learning in the future.

关 键 词：镱原子光钟 FPGA控制系统 数据处理 机器学习 

分 类 号：O431[机械工程—光学工程]