用于谱线检测的随机解调器硬件设计与FPGA实现

发布时间：2018-06-26 13:20

本文选题：谱线检测 + 稀疏信号　；参考：《云南大学》2015年硕士论文

【摘要】：谱线检测是射电天文观测的关键技术之一。根据观察到的星体谱线,可以了解星体的一些重要特征,比如运行轨迹、总粒子数密度、年龄等。这些特征所对应的分子辐射谱线载频位置极高,都在GHz量级以上。基于传统高速数据采集的射电天文终端设备存在很多局限性,无法很好地满足天文学家同时观测更多谱线和同时观测频率相差很远的分子多个转动跃迁的要求。射电天文信号是在频域上稀疏的信号。压缩感知理论正是针对在某个域上稀疏的信号提出的,该理论可以解决传统采样在高频信号采集中ADC采样率高和数据存储容量大的问题。基于压缩感知理论的信号采集把采样过程和压缩过程合二为一,以较低的采样率采样得到较少的数据,最后通过求解最优问题来对原始稀疏信号进行重构。随机解调器、多倍集采样和调制宽带转换器三种压缩采样模型的提出使得压缩感知理论进入实用阶段。论文对随机解调器信号采样与重构理论进行分析研究,在仿真实验的基础上,设计了硬件电路,并通过FPGA实现随机解调器采样与重构。首先,论文介绍了压缩感知的基本原理,包括压缩感知理论的三大核心问题以及三种压缩采样模型。其次,论文深入分析了随机解调器信号采样与重构方法,并进行了算法仿真实验。最后,设计了硬件电路实现随机解调器。硬件电路包括如下几个部分：信号预处理部分通过精心挑选合适的芯片并对每个模块进行仿真验证后设计制作了相应的电路板；信号采样与数据缓存通过FPGA对ADC和SDRAM控制完成；原始信号的稀疏矢量重构利用OMP算法,首先对OMP算进行改进,然后在设计中通过资源复用、并行计算和避开复杂运算(如开方运算)等方法在FPGA上优化设计实现,既加快了运算速度又减少硬件电路的复杂度。对随机解调器的仿真验证了系统的可行性。硬件实现的随机解调器系统充分利用了FPGA并行计算的优势,能以较低的采样率对频域稀疏信号进行采样并对其进行谱线观测。实验结果表明,该系统实现了压缩采样,以40MHz的采样率成功的对最高频率为80MHz的信号进行谱线检测。
[Abstract]:Spectral line detection is one of the key techniques in radio astronomical observation. According to the observed spectral lines of the stars, some important characteristics of the stars can be understood, such as trajectory, total particle density, age and so on. These characteristics correspond to the extremely high carrier frequency positions of the molecular radiation lines, all of which are of the order of magnitude above GHz. The radio astronomical terminal equipment based on the traditional high-speed data acquisition has many limitations, which can not meet the requirements of the astronomer to observe more spectral lines at the same time and to observe many rotational transitions of molecules with far different frequency simultaneously. Radio astronomical signals are sparse signals in frequency domain. Compression sensing theory is proposed for sparse signals in a certain domain. This theory can solve the problems of high ADC sampling rate and large data storage capacity in traditional sampling in high frequency signal acquisition. The signal acquisition based on compression sensing theory combines the sampling process with the compression process, and obtains less data at a lower sampling rate. Finally, the original sparse signal is reconstructed by solving the optimal problem. Three compression sampling models, random demodulator, multi-fold sampling and modulation wideband converter, have brought the theory of compression sensing into practical stage. In this paper, the theory of signal sampling and reconstruction of random demodulator is analyzed and studied. On the basis of simulation experiment, the hardware circuit is designed, and the sampling and reconstruction of random demodulator is realized by FPGA. Firstly, the basic principle of compression sensing is introduced, including three core problems of compression sensing theory and three compression sampling models. Secondly, the method of sampling and reconstruction of random demodulator signal is deeply analyzed, and the algorithm simulation experiment is carried out. Finally, the hardware circuit is designed to realize the random demodulator. The hardware circuit includes the following parts: the signal preprocessing part designs and manufactures the corresponding circuit board after selecting the appropriate chip carefully and carries on the simulation verification to each module, the signal sampling and the data cache control the ADC and the SDRAM through the FPGA to complete; The sparse vector reconstruction of the original signal uses OMP algorithm to improve the OMP calculation first, and then optimizes the design on FPGA by means of resource reuse, parallel computing and avoiding complex operation (such as square operation). It not only speeds up the operation speed but also reduces the complexity of the hardware circuit. The simulation of the random demodulator verifies the feasibility of the system. The hardware implementation of the stochastic demodulator system makes full use of the advantages of FPGA parallel computing, and can sample the sparse signals in frequency domain at a low sampling rate and observe the spectral lines. The experimental results show that the system realizes the compression sampling and successfully detects the spectral lines of the signals with the highest frequency of 80MHz at the sampling rate of 40MHz.
【学位授予单位】：云南大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TN763

【参考文献】