时空二维侦察处理技术
|
摘要
宽频段高分辨率时-空二维谱估计是当前阵列信号处理研究的一个重要方向，虽然在众多学者的努力下已取得了丰富的理论成果，但是一直以来宽频段阵列系统的工程实现存在很多困难。近年来随着微电子技术的发展，以及高性能时空二维谱估计算法的出现，使宽频段高分辨率时-空二维谱估计准实时系统的硬件实现成为可能。
在本实验系统中，我们设计了一个5阵元的L型非均匀阵列天线，将的频段划分成的子频段，然后将每个子频段混频到的处理基带后再进行的同步欠采样。其中有一个阵元输出有一通道经延时后再采样。对采样数据进行FFT和谱峰搜索，截取谱峰附近数据构造一路信号的四阶累积量矩阵和二路信号的互四阶累积量矩阵，构成了一对ESPRIT算法所需要的旋转不变矩阵束。利用特征值相位和基于整数搜索法来实现信号频率和二维到达角的无模糊估计。
本文主要介绍了基于四阶累积量的时-空二维谱估计算法，并以此为基础来实现时频空三维侦察机；宽频段阵列系统各种误差参数的校正算法，实验系统的硬件驱动程序的开发和应用软件的开发。系统采用当前比较先进的DSP技术和PCI局部总线技术，样机由三块FFT卡和一块主处理卡组成，其中FFT卡选用的是TMS320C31 DSP，而主处理卡采用的是TMS320C67 DSP。系统选用工控机为控制平台，Windows为操作系统。为了实现硬件卡之间、应用程序和硬件卡之间的数据传输，需要开发VxD驱动程序，本文详细介绍了驱动程序的开发过程。在阵列信号处理中，阵列系统的各种误差对算法性能的影响是比较大的，所以本文也对阵列的各种误差校正算法进行了研究。
该样机系统还属于实验性研制阶段，但通过模拟试验结果表明，该处理机可对时空均欠采样的数字信号进行频率和二维到达角的无模糊估计，并且性能达到预期设计的要求。虽然该处理机是用来处理平稳窄带信号的，但是该实验样机为非平稳信号的参数估计系统的研制提供了许多理论和工程经验，也为高分辨率时-空二维谱估计技术的实际应用打下了坚实的基础。
The high-resolution spatio-temporal two-dimensional (2-D) spectra estimation of wide-band signal acts an important role in modern array signal processing. Though abundant results acquired in theory, there are a lot of difficulties in the real time application of wide-band system. The development of micro-electronic and high performance algorithms of spatio-temporal 2-D spectra estimation offered probability for us to apply it on a real time system now.
In prototype system, a L-shaped non-uniform linear array with 5 elements is utilized to receive spatial signal with carrier frequency among 2-18GHz. The entire frequency band is decomposed into 1GHz sub-bands and then is mixed to 1GHz base band. The base band output of each antenna is sent through a synchronous 250MHz undersampling channels. One antenna has a delayed 0.5ns relative to the others (prior to sampling). FFT is applied to data and spectra peak search is done, then using the data near the spectra peaks to compute two 4th-order cumulants matrix to form rotational invariance matrix pencil which the ESPRIT algorithm can be applied to. The non-aliasing frequency and non-ambiguous 2-D angles is obtained by the phase characteristic of eignevalue and integral search method
In this dissertation, an algorithm of spatio-temporal 2-Dspectra estimation based on 4th-order cumulants was introduced and used to realize a prototype scout, a calibration algorithm in the wide-band array system is proposed. It also presents the development of hardware driver and application software for the prototype system. The technique of DSP and PCI local bus were applied in the system. Three FFT boards (TMS320C31 DSP) and one main processing board(TMS320C67 DSP) contributed to the system. We used the special computer as the controlling platform and Windows as the OS. The development of VxD driver software is describe in detail which is used to realize data transferring between hardware and application software, also data transferring among all boards. In array signal processing, various kinds of errors can degrade the performance of signal processing algorithm, errors calibration algorithm is also studied in this dissertation.
Now our prototype system is in experiment, but the simulation result show that it can do non-alias frequency and 2D-DOAs of spatio-temporal undersampling digital signal and obtain the aimed performance. Though out system is designed to process stable narrow-band signal, it also provide experience for study of unstable signal parameter estimation system in theory and engineering, and establish a basement for the application of high-resolution spato-temporat 2-D spectra setimation.
在本实验系统中，我们设计了一个5阵元的L型非均匀阵列天线，将的频段划分成的子频段，然后将每个子频段混频到的处理基带后再进行的同步欠采样。其中有一个阵元输出有一通道经延时后再采样。对采样数据进行FFT和谱峰搜索，截取谱峰附近数据构造一路信号的四阶累积量矩阵和二路信号的互四阶累积量矩阵，构成了一对ESPRIT算法所需要的旋转不变矩阵束。利用特征值相位和基于整数搜索法来实现信号频率和二维到达角的无模糊估计。
本文主要介绍了基于四阶累积量的时-空二维谱估计算法，并以此为基础来实现时频空三维侦察机；宽频段阵列系统各种误差参数的校正算法，实验系统的硬件驱动程序的开发和应用软件的开发。系统采用当前比较先进的DSP技术和PCI局部总线技术，样机由三块FFT卡和一块主处理卡组成，其中FFT卡选用的是TMS320C31 DSP，而主处理卡采用的是TMS320C67 DSP。系统选用工控机为控制平台，Windows为操作系统。为了实现硬件卡之间、应用程序和硬件卡之间的数据传输，需要开发VxD驱动程序，本文详细介绍了驱动程序的开发过程。在阵列信号处理中，阵列系统的各种误差对算法性能的影响是比较大的，所以本文也对阵列的各种误差校正算法进行了研究。
该样机系统还属于实验性研制阶段，但通过模拟试验结果表明，该处理机可对时空均欠采样的数字信号进行频率和二维到达角的无模糊估计，并且性能达到预期设计的要求。虽然该处理机是用来处理平稳窄带信号的，但是该实验样机为非平稳信号的参数估计系统的研制提供了许多理论和工程经验，也为高分辨率时-空二维谱估计技术的实际应用打下了坚实的基础。
The high-resolution spatio-temporal two-dimensional (2-D) spectra estimation of wide-band signal acts an important role in modern array signal processing. Though abundant results acquired in theory, there are a lot of difficulties in the real time application of wide-band system. The development of micro-electronic and high performance algorithms of spatio-temporal 2-D spectra estimation offered probability for us to apply it on a real time system now.
In prototype system, a L-shaped non-uniform linear array with 5 elements is utilized to receive spatial signal with carrier frequency among 2-18GHz. The entire frequency band is decomposed into 1GHz sub-bands and then is mixed to 1GHz base band. The base band output of each antenna is sent through a synchronous 250MHz undersampling channels. One antenna has a delayed 0.5ns relative to the others (prior to sampling). FFT is applied to data and spectra peak search is done, then using the data near the spectra peaks to compute two 4th-order cumulants matrix to form rotational invariance matrix pencil which the ESPRIT algorithm can be applied to. The non-aliasing frequency and non-ambiguous 2-D angles is obtained by the phase characteristic of eignevalue and integral search method
In this dissertation, an algorithm of spatio-temporal 2-Dspectra estimation based on 4th-order cumulants was introduced and used to realize a prototype scout, a calibration algorithm in the wide-band array system is proposed. It also presents the development of hardware driver and application software for the prototype system. The technique of DSP and PCI local bus were applied in the system. Three FFT boards (TMS320C31 DSP) and one main processing board(TMS320C67 DSP) contributed to the system. We used the special computer as the controlling platform and Windows as the OS. The development of VxD driver software is describe in detail which is used to realize data transferring between hardware and application software, also data transferring among all boards. In array signal processing, various kinds of errors can degrade the performance of signal processing algorithm, errors calibration algorithm is also studied in this dissertation.
Now our prototype system is in experiment, but the simulation result show that it can do non-alias frequency and 2D-DOAs of spatio-temporal undersampling digital signal and obtain the aimed performance. Though out system is designed to process stable narrow-band signal, it also provide experience for study of unstable signal parameter estimation system in theory and engineering, and establish a basement for the application of high-resolution spato-temporat 2-D spectra setimation.
引文
[1]M.Wax,T.-J.Shan and T.Kailath,"Spatio-Temporal Spectral Analysis by Eigenstructure Methods" , IEEE Trans. ASSP ,Vol. ASSP-32, No. 4, pp. 817～827, Aug.1984.
[2]L.-H.Zou and L.Yin,"Spatio-Temporal Spectral Analysis by SVD of Signal-Matrix" , Proc.IEEE,ICASSP-87,pp.2332～2335,April,1987.
[3]Q.-Y.Yin and L.-H.Zou,“Separately Estimating the Frequency-Wavenumber of Coherent Sources by Signal Eigevector Smoothing Technique" ,Proc.IEEE,ICASSP-88,pp.2905～2908,1988.
[4]W.S.McCormick and J.B.Y.Tsui,“Suboptimal Real-Time Frequency/Incident-Angle Estimator for Mutiple Radar Pulses" ,IEEE Proceedings-F,Vol.3,pp.247～254,June 1991.
[5]M.D.Zoltowski,C.P.Mathews,“Real-Time Requency and 2-D Angle Estimation with Sub-Nyquist Spatio-Temporal Sampling" ,IEEE Trans.SP,Vol.42,No.10,pp.2781～2794,Oct.1994.
[6]M.Harrdt,J.A.Nossek,“3-D unitary ESPRIT for joint 2-D angle and carrier frequency estimation",IEEE ICASSP,(1):255～258,1997.
[7]M.Viberg,P.Stoica,“A computationally efficient method for joint direction finding and frequency estimation in colored noise",32nd Asilomar Conference on Signals,Systems and Computers,(2):1547～1551,1998.
[8]J.W.C.Robinson,“Multiple invariance spatio-temporal spectral estimation",Scientific report,NTIS No:PB99-102345,XAB.
[9]L.Ge,T.-Q.Chen,X.-F.Huang,“Simultaneous Frequency and Direction Estimation form Parallel-Array Data",IEEE Proc. -Radar,Sonar,Navg, Vol.142,No.1,pp.6～10,Feb.1995.
[10]M.D.Zoltowski and D.Strvrindes,“Sensor Array Signal Processing via a Procrusters Rotations Based Eigenanalysis of the ESPRIT Data Pencil",IEEE Trans.ASSP,Vol.37,pp.832～861,No.6,June 1989.
[11]G.H.Xu,S.D.Silverstein,T.Kailath,“beamspace ESPRIT",IEEE Trans.SP, Vol.42,No.2, pp.349～355,Feb.1994.
[12]G.H.Xu,S.D.Silverstein,T.Kailath,“Parallel inplementation and performance analysis of beamspace ESPRIT",Proc.IEEE ICASSP, pp.1497～1500,Apr.1991.
[13]A.L.Swindlehurst."Detection and estimation in the presence of signals with uncalibrated spatial signature",Sig.Proc.April.1996.
[14]Y.Rockah,P.M.Schultheiss,“Array shape calibration using source in unknown locations-PartⅠ:Far-field sources",IEEE Trans ASSP-35(3),pp.286～299,Mar.1987.
[15]M.C.Dogan,J.M.Mendel,“Application of cumulants to array processing-Part Ⅰ:Aperture extension and array calibration”,IEEE Trans, SP43(5):1200～1216,May 1995.
[16]T.H.Liu,J.M.Mendel,“Application of cumulants to array processing-Part Ⅴ:Sensitivity Issues",IEEE Trans.SP 47(3):746～759,Mar.1999.
[17] Y.Rockah,P.M.Schultheiss,“Array shape calibration using source in unknown locations-Part
Ⅱ:Near-field sources and estimator implementation",IEEE Trans ASSP-35(6):724～735,June 1987.
[18]J.T.Lo,S.L.Marple,“Eigenstructure method for array sensor location",Proc.IEEE ICASSP:2260～2263,Texas,Dallas,1987.
[19]A.L.Swindlehurst,T.Kailath,“A performance analysis of subspace-based methods in the presence of model errors,PartⅠ:the MUSIC algorithm",IEEE Trans.SP-40(7):1758～1774,July 1992.
[20]A.J.Weiss,B.Friedlander,“Array shape calibration using sources in unknown locations—A maximum likelihood approach"'IEEE Trans. Acout.,Speech,Signal Processing,Vol37, pp.1958～1966,Dec.1989.
[21C.M.S.See,“Method for array calibration in high-resolution sensor array processing",IEE Proc.RSN-(142)3,June,1995.
[22]I.S.D.Solomon,D.A.Gray,Y.I.Abramovich,“Receiver array calicbration using disparate Sources",IEEE Trans.Vol.47,N0.3,pp.496～505, Mar.1999.
[23]M.Viberg,B.Ottersten,T.Kailath,“Detection and estimation in sensor arrays using weighted subspace fitting",IEEE Trans.Signal Processing, Vol.39,pp.2436～2449,Nov.1991.
[25]张贤达著，《现代信号处理》，清华大学出版社，1995年10月第1版。
[26]葛利嘉，“高分辨时空二维谱估计方研究”，博士论文，电子科技大学，1994年4月。
[27]王激扬，“测频测向一体化新体制雷达侦察接收机研制——信号处理机”，博士论文，电子科技大学，2000年10月。
[28]黄佑勇，“基于高阶累积量的阵列信号多参数估计技术”，博士论文，电子科技大学，2001年1 0月。
[29]刘德树，罗景青和张剑去编著，《空间谱估计技术及其应用》，中国科技大学出版社，1997年8月第1
[30]陈天麒，“宽频段时空二维谱估计中的关键技术”，电子科技大学研究报告
[31]王激扬，黄估勇，陈天麒，“空间欠采样信号DOA估计的阵列配置及解模糊算法”，电波科学学报，Vol.14,No.4,pp.450～456，Dec.1998.
[32]艾德才等编著 《Pentium/80486实用汇编语言程序设计》，清华大学出版社，1997年11月第1版。
[33]彭礼孝编著，《虚拟设备驱动程序开发起步与进阶》，人民邮电出版社，2000年12月第1版。
[34]Applied Micro Circuits Corporation,《S5920 PCI Target Interface Data Book》，Copyright(c)1997.
[2]L.-H.Zou and L.Yin,"Spatio-Temporal Spectral Analysis by SVD of Signal-Matrix" , Proc.IEEE,ICASSP-87,pp.2332～2335,April,1987.
[3]Q.-Y.Yin and L.-H.Zou,“Separately Estimating the Frequency-Wavenumber of Coherent Sources by Signal Eigevector Smoothing Technique" ,Proc.IEEE,ICASSP-88,pp.2905～2908,1988.
[4]W.S.McCormick and J.B.Y.Tsui,“Suboptimal Real-Time Frequency/Incident-Angle Estimator for Mutiple Radar Pulses" ,IEEE Proceedings-F,Vol.3,pp.247～254,June 1991.
[5]M.D.Zoltowski,C.P.Mathews,“Real-Time Requency and 2-D Angle Estimation with Sub-Nyquist Spatio-Temporal Sampling" ,IEEE Trans.SP,Vol.42,No.10,pp.2781～2794,Oct.1994.
[6]M.Harrdt,J.A.Nossek,“3-D unitary ESPRIT for joint 2-D angle and carrier frequency estimation",IEEE ICASSP,(1):255～258,1997.
[7]M.Viberg,P.Stoica,“A computationally efficient method for joint direction finding and frequency estimation in colored noise",32nd Asilomar Conference on Signals,Systems and Computers,(2):1547～1551,1998.
[8]J.W.C.Robinson,“Multiple invariance spatio-temporal spectral estimation",Scientific report,NTIS No:PB99-102345,XAB.
[9]L.Ge,T.-Q.Chen,X.-F.Huang,“Simultaneous Frequency and Direction Estimation form Parallel-Array Data",IEEE Proc. -Radar,Sonar,Navg, Vol.142,No.1,pp.6～10,Feb.1995.
[10]M.D.Zoltowski and D.Strvrindes,“Sensor Array Signal Processing via a Procrusters Rotations Based Eigenanalysis of the ESPRIT Data Pencil",IEEE Trans.ASSP,Vol.37,pp.832～861,No.6,June 1989.
[11]G.H.Xu,S.D.Silverstein,T.Kailath,“beamspace ESPRIT",IEEE Trans.SP, Vol.42,No.2, pp.349～355,Feb.1994.
[12]G.H.Xu,S.D.Silverstein,T.Kailath,“Parallel inplementation and performance analysis of beamspace ESPRIT",Proc.IEEE ICASSP, pp.1497～1500,Apr.1991.
[13]A.L.Swindlehurst."Detection and estimation in the presence of signals with uncalibrated spatial signature",Sig.Proc.April.1996.
[14]Y.Rockah,P.M.Schultheiss,“Array shape calibration using source in unknown locations-PartⅠ:Far-field sources",IEEE Trans ASSP-35(3),pp.286～299,Mar.1987.
[15]M.C.Dogan,J.M.Mendel,“Application of cumulants to array processing-Part Ⅰ:Aperture extension and array calibration”,IEEE Trans, SP43(5):1200～1216,May 1995.
[16]T.H.Liu,J.M.Mendel,“Application of cumulants to array processing-Part Ⅴ:Sensitivity Issues",IEEE Trans.SP 47(3):746～759,Mar.1999.
[17] Y.Rockah,P.M.Schultheiss,“Array shape calibration using source in unknown locations-Part
Ⅱ:Near-field sources and estimator implementation",IEEE Trans ASSP-35(6):724～735,June 1987.
[18]J.T.Lo,S.L.Marple,“Eigenstructure method for array sensor location",Proc.IEEE ICASSP:2260～2263,Texas,Dallas,1987.
[19]A.L.Swindlehurst,T.Kailath,“A performance analysis of subspace-based methods in the presence of model errors,PartⅠ:the MUSIC algorithm",IEEE Trans.SP-40(7):1758～1774,July 1992.
[20]A.J.Weiss,B.Friedlander,“Array shape calibration using sources in unknown locations—A maximum likelihood approach"'IEEE Trans. Acout.,Speech,Signal Processing,Vol37, pp.1958～1966,Dec.1989.
[21C.M.S.See,“Method for array calibration in high-resolution sensor array processing",IEE Proc.RSN-(142)3,June,1995.
[22]I.S.D.Solomon,D.A.Gray,Y.I.Abramovich,“Receiver array calicbration using disparate Sources",IEEE Trans.Vol.47,N0.3,pp.496～505, Mar.1999.
[23]M.Viberg,B.Ottersten,T.Kailath,“Detection and estimation in sensor arrays using weighted subspace fitting",IEEE Trans.Signal Processing, Vol.39,pp.2436～2449,Nov.1991.
[25]张贤达著，《现代信号处理》，清华大学出版社，1995年10月第1版。
[26]葛利嘉，“高分辨时空二维谱估计方研究”，博士论文，电子科技大学，1994年4月。
[27]王激扬，“测频测向一体化新体制雷达侦察接收机研制——信号处理机”，博士论文，电子科技大学，2000年10月。
[28]黄佑勇，“基于高阶累积量的阵列信号多参数估计技术”，博士论文，电子科技大学，2001年1 0月。
[29]刘德树，罗景青和张剑去编著，《空间谱估计技术及其应用》，中国科技大学出版社，1997年8月第1
[30]陈天麒，“宽频段时空二维谱估计中的关键技术”，电子科技大学研究报告
[31]王激扬，黄估勇，陈天麒，“空间欠采样信号DOA估计的阵列配置及解模糊算法”，电波科学学报，Vol.14,No.4,pp.450～456，Dec.1998.
[32]艾德才等编著 《Pentium/80486实用汇编语言程序设计》，清华大学出版社，1997年11月第1版。
[33]彭礼孝编著，《虚拟设备驱动程序开发起步与进阶》，人民邮电出版社，2000年12月第1版。
[34]Applied Micro Circuits Corporation,《S5920 PCI Target Interface Data Book》，Copyright(c)1997.