时间交替采集系统通道失配误差校准及带宽扩展方法研究

发布时间：2018-03-09 20:21

本文选题：时间交替　切入点：模数转换　出处：《国防科学技术大学》2015年博士论文　论文类型：学位论文

【摘要】：数据采集设备作为混合信号处理系统中的重要组成单元,是连接真实的模拟世界和虚拟的数字世界之间的唯一桥梁,是支撑现代信号处理的基石。随着测量频谱宽度和瞬时动态范围的不断提升,对数据采集设备的采样速度和采样精度提出了更高的要求。但受到制备工艺和制造材料的限制,单个模数转换器(Analog-to-Digital Converter,ADC)的速度和精度已无法满足日益增长的应用需求。在现有芯片的条件下,如何进一步提升数据采集系统的性能指标成为一个亟待解决的问题。时间交替模数转换(Time-Interleaved ADCs,TI-ADCs)系统利用多个低速ADC并行交错采样能够将采样速度以通道数量等比例提升,突破现有芯片性能的制约。然而,TI-ADCs系统的动态性能对通道间失配误差极为敏感,任何微小的失配都会引起严重的动态性能恶化。此外TI-ADCs的另一个局限是其模拟带宽并不能如采样速率一样随通道数量成比例提升,而是受单个ADC的限制。本文针对TI-ADCs系统中通道失配误差的估计和补偿问题以及模拟带宽的数字扩展技术展开研究,主要工作如下:第2章研究了TI-ADCs系统模型和失配误差影响,推导并验证了失配误差强度和系统动态性能之间的定量关系,此结果能够为TI-ADCs系统确定误差容限,为系统设计和非线性失配校准算法提供精度参考标准。第3章研究了TI-ADCs非线性失配误差的校准方法,分别提出了一种前向校准算法和一种自适应盲校准算法。其中,前向校准算法利用训练信号估计失配参数进而利用数字级联补偿结构优化系统性能,仿真表明,算法能够有效抑制非线性失配谐波对系统动态性能的影响。该算法在达到相同校准精度的前提下,相比将单通道ADC非线性误差优化算法简单扩展至M通道TI-ADCs显著降低了计算复杂度。其次,提出了一种自适应盲校准算法以满足对时变非线性失配误差的优化,算法借助于一定比例的过采样以获取失配谐波分量,利用最小均方算法完成对失配参数的自适应跟踪。算法利用离散傅里叶级数将M周期的非线性失配误差转换为M个与时间不相关的失配系数,从而能够显著提升模拟频带利用率。利用失配镜像之间的复共轭关系对M通道的自适应校准结构进行了优化。通过仿真验证了算法在不同输入信号和不同失配强度下的校准性能。第4章针对线性失配误差与非线性失配误差共存的情况,提出并验证了一种联合盲校准算法。该算法利用一定比例的过采样和自适应滤波结构实现混合失配误差的自适应校准,仿真结果表明,相比利用现有算法处理线性与非线性失配,该算法能够实现更显著的动态性能提升。针对宽带TI-ADCs系统中失配误差与输入频率的相关性,进一步将联合校准算法扩展至频率相关的混合失配应用场景。通过仿真验证了算法在不同失配模型,输入信号以及失配强度下的对混合失配误差的优化能力。第5章提出了一种基于最小二乘FIR滤波器的带宽扩展方法,首次提出并验证了最小二乘准则下的滤波器阶数估计函数,在此基础上给出了详细的滤波器设计参数配置方法,从而能够有效地缩短设计时间,并为顶层系统设计提供准确的资源配置参考。针对不同应用场景的需求,本文分别提出了优先噪声抑制和优先杂散抑制的滤波器的阶数估计函数和设计参数配置方法,利用提出的估计函数和设计参数配置方法能够根据给定的给定的带宽扩展参数以及通带和噪声损耗的约束条件,准确快速地逼近滤波器设计的最优参数,如最小滤波器阶数,权值比重,过渡带宽度等,从而能够有效地减少滤波器设计时间。最后,通过仿真验证了两种应用场景下阶数估计函数的精度以及设计参数配置方法的性能。
[Abstract]:The data acquisition equipment as an important unit in mixed signal processing system, is the only bridge between the real world and the virtual simulation of the digital world, is the support of modern signal processing. With the measurement of the width of the spectrum and the cornerstone of the instantaneous dynamic range of continuous improvement, put forward higher requirements for data acquisition device and sampling speed the precision of sampling. But by the preparation process and manufacturing material limit, a single analog to digital converter (Analog-to-Digital Converter, ADC) of the speed and accuracy have been unable to satisfy the application demand. In the existing chip conditions, how to further enhance the performance of the data acquisition system has become an urgent problem in time interleaved analog-to-digital conversion. (Time-Interleaved ADCs TI-ADCs) system using a plurality of low-speed ADC parallel interleaved sampling can be sampled in proportion to the number of channels such as lifting speed L, which break through the existing chip performance. However, the dynamic performance of the TI-ADCs system is very sensitive to the mismatch error, any slight mismatch will cause serious deterioration of dynamic performance. Another limitation of TI-ADCs is the analog bandwidth and not as a kind of sampling rate with the number of channels in proportion to improve but, by a single ADC. According to the estimation and compensation of channel mismatch error in TI-ADCs system and digital analog bandwidth expansion technologies are studied, the main work is as follows: the second chapter studies the TI-ADCs system model and the mismatch effect, deduced and verified the quantitative relationship between the mismatch strength and dynamic performance of the system the results, to determine the tolerance for the TI-ADCs system, provide a reference standard for the accuracy of the system design and nonlinear mismatch calibration algorithm. The third chapter studies the TI-ADCs nonlinear mismatch error The calibration method of difference, respectively, this paper proposes a forward calibration algorithm and an adaptive quasi blind algorithm. The algorithm uses forward calibration parameter mismatch and the use of digital cascade compensation structure optimization of the system performance, the training signal estimation. The simulation results show that the algorithm can effectively restrain the effect of nonlinear harmonic mismatch on the dynamic performance of the system. On the premise of achieving the same algorithm precision, compared to the single channel ADC nonlinear error optimization algorithm is simple extension to M channel TI-ADCs significantly reduces the computational complexity. Secondly, this paper proposed an adaptive blind calibration algorithm to meet the optimal time-varying and nonlinear mismatch error, over sampling to obtain the algorithm based on harmonic component mismatch a certain proportion of the least mean square algorithm to track the adaptive parameter mismatch. The proposed algorithm uses discrete Fourier series with nonlinear error of M cycle loss Conversion to M is not related to the mismatch of coefficient and time, which can significantly improve the analog bandwidth utilization. Optimized by complex conjugate relationship between image with adaptive calibration structure of M algorithm in different channel. The input signal and the different intensity of the mismatch calibration performance is verified by simulation. The fourth chapter according to the linear mismatch error and nonlinear mismatch error coexist situation, propose and validate a joint blind calibration algorithm. Over sampling and adaptive filtering structure to realize hybrid mismatch error calibration algorithm using the adaptive proportion. The simulation results show that compared with the algorithm of linear and nonlinear mismatch existing, the algorithm can to realize the dynamic performance significantly improved. According to the correlation in wideband TI-ADCs system mismatch error and input frequency, further joint calibration algorithm extended to frequency mixing Combination of mismatch scenarios. The algorithm in different mismatch model is verified by simulation, and the input signal strength mismatch on the mixed mismatch error optimization. The fifth chapter puts forward a method of extending the bandwidth of the filter based on least squares FIR, first proposed and verified the function estimation of filter order least squares criterion on this basis, gives the design parameters of filter configuration method in detail, which can effectively shorten the design time, the allocation of resources and provide accurate reference for the top-level system design. According to different application scenarios, this paper puts forward design function and parameter configuration method to estimate the order of priority priority filter noise and spurious suppression the function and design parameters is estimated by using the proposed method can according to the given bandwidth expansion parameters and band noise and loss about Beam conditions, accurate approximation of optimal filter design parameters, such as the minimum filter order, weight proportion, transition zone width, which can effectively reduce the filter design time. Finally, through the simulation of two kinds of application scenarios order estimation precision of the performance function and design parameter configuration method.

【学位授予单位】：国防科学技术大学
【学位级别】：博士
【学位授予年份】：2015
【分类号】：TN792

【相似文献】