Research And Implementation Of Correction Algorithm For Ultra-Wideband Acquisition System Based On Subband Decomposition

With the rapid development of digital signal theory and related processing technology,digital signal processing system has become the mainstream choice of modern electronic systems to achieve a variety of functions.Among them,data acquisition system has become the key conversion component between analog electrical signal and digital signal.As the core component,analog-to-digital converter(ADC)plays a decisive role in the performance of this data acquisition system.At present,the existing large-scale commercial ADC monolithic performance indicators are still difficult to meet the requirements of Ultra-Wideband(UWB)data acquisition system.Parallel sampling technology must be used to improve system performance.The basic idea of parallel sampling technology is to form a data acquisition system equivalent to high performance ADC by using multiple ADCs with low performance indicators to form a sampling array according to certain rules.Traditional parallel sampling techniques focus on the improvement of sampling rate,which can not meet the bandwidth requirements of UWB Band System.Since 2007,the "Mixing + FI-ADC" parallel sampling structure represented by Digital Bandwidth-Interleaved(DBI)of Le Croy has been applied to Ultra-Wide Bandwidth Data Acquisition System.This kind of structure utilizes the feature of FI-ADC sub-band decomposition sampling and reduces the actual signal bandwidth of the sampling path into a single ADC by mixing,making the overall system bandwidth increase as much as the sampling rate.In these structures,DBI technology has the advantages of simple principle,easy to understand,high clock robustness,low requirement on components,no dependence on structural design to reduce errors,and is suitable for system-level engineering applications.Based on the above research background,this paper focuses on various issues in the design and implementation of digital bandwidth alternation technology.The specific research contents are as follows:(1)Based on the theory of FI-ADC,filter bank and multi rate signal processing,the mathematical model of DBI technology is established,and the conditions for accurate reconstruction of DBI are given,which provides a theoretical basis for subsequent research.The errors affecting the accurate reconstruction of DBI system are summarized into four types and seven subtypes.Using the above classification,the overall correction problem of complex DBI system is transformed into the correction problem of 7 types of errors,which will be used as the direction of follow-up research.(2)The mirror frequency signal of analog mixing is considered together with the spurious signal caused by the non-ideal of analog mixing.The conditions for filtering analog mixed-frequency stray signals are obtained under the principle of uniform ADC data acquisition channel design.The sub-band division and local frequency selection algorithm for DBI system are obtained.This algorithm can filter out the analog mixing spurious signals in all the sub-bands except sub-band 2 of the DBI system,and only the third-order cross-modulation difference signals remain in the sub-band 2.The error handling between the mirror frequency signal of analog mixing and the non-ideal of analog mixing is achieved.(3)The mirror frequency signal of digital mixing is considered together with the interpolated mixing signal.The estimation formula for the upper limit of the spurious signal and the algorithm for suppressing and estimating the digital mixed spurious signal are obtained.Error handling of mirrored frequency signals and interpolated overlapped signals for digital mixing is implemented.(4)For the application scenario of UWB data acquisition,the conditions for accurate DBI rebuilding and the performance evaluation index of UWB data acquisition system are analyzed,the constraints for approximate and accurate reconstruction of distortion function are improved,and the difficulties for subsequent correction are reduced.An amplitude correction filter design algorithm based on quasi-minimum residual method and a weighted least squares all-pass filter design algorithm based on complex cepstrum are presented.The amplitude and phase approximation PR of the distortion function are achieved,respectively,and the approximation PR of the system is finally achieved.(5)For the overlapping signal of adjacent sub-bands caused by the decomposition and merging of common factor bands in the "Mixing + FI-ADC" structure such as DBI technology,the influence of overlapping signal overlapping is analyzed,and the overlapping correction algorithm which does not require additional resources is presented in combination with the structure characteristics of the data acquisition system using the "Mixing + FI-ADC" parallel sampling technology.The correction of the overlapping band problem of the "Mixing + FI-ADC" structure is implemented.(6)Analyzing the problems faced by UWB data acquisition system testing at present,using the features of "Mixing + FI-ADC" structure such as DBI,this dissertation puts forward the data acquisition system testing method for this kind of structure.Accurate and convenient measurement of UWB data acquisition system is achieved.For all electronic systems that require ultra-wideband data acquisition systems,the DBI technology studied in this dissertation can be used to increase the sampling rate and bandwidth of the system.In addition,since the DBI technology is the most basic "Mixing+ FI-ADC" structure in principle,other structures can be regarded as improvements to the digital bandwidth alternation technology.The related research results of this thesis can also be applied when using other data acquisition systems with the structure of mixing and FIADC.