張學(xué)敬 楊志偉 廖桂生
?
共形陣列幅相誤差校正快速算法
張學(xué)敬*楊志偉 廖桂生
(西安電子科技大學(xué)雷達(dá)信號(hào)處理國(guó)家重點(diǎn)實(shí)驗(yàn)室 西安 710071)
基于子空間的聯(lián)合迭代算法可以實(shí)現(xiàn)對(duì)空間信源方位和陣列幅相誤差參數(shù)的聯(lián)合估計(jì)。當(dāng)對(duì)共形陣列進(jìn)行幅相誤差校正時(shí),由于其空域?qū)蚴噶坎痪哂蠽ander monde結(jié)構(gòu),導(dǎo)致快速高分辨空間譜估計(jì)方法無(wú)法直接應(yīng)用,而利用2維譜峰搜索實(shí)現(xiàn)空間方位估計(jì)的運(yùn)算量較大,限制了算法在共形陣列上的應(yīng)用。針對(duì)此問(wèn)題,該文提出一種借助虛擬陣列實(shí)現(xiàn)共形陣列幅相誤差校正的新方法。該方法利用虛擬陣列的特殊結(jié)構(gòu)快速實(shí)現(xiàn)對(duì)信源的DOA估計(jì),省去了譜峰搜索過(guò)程,因而運(yùn)算復(fù)雜度低,便于工程實(shí)現(xiàn)。理論分析和仿真結(jié)果驗(yàn)證了所提算法的有效性,可為共形陣列的工程應(yīng)用提供參考。
DOA估計(jì);幅相誤差校正;共形陣列;虛擬內(nèi)插


本文通過(guò)深入分析WF算法,將虛擬內(nèi)插變換法[14]引入到共形陣列的幅相誤差校正中。算法通過(guò)調(diào)整虛擬陣列變換矩陣,對(duì)虛擬子空間進(jìn)行修正,利用修正后的虛擬子空間之間的正交關(guān)系實(shí)現(xiàn)對(duì)信源方位的DOA估計(jì),并由此得到原共形陣列的幅相誤差參數(shù)。該算法無(wú)需譜峰搜索,具有較低的運(yùn)算復(fù)雜度,可以快速高效地完成共形陣列幅相誤差和信源方位的聯(lián)合估計(jì)。文中對(duì)算法進(jìn)行了理論分析和仿真驗(yàn)證,可為共形陣列的應(yīng)用提供參考。





圖1 任意幾何結(jié)構(gòu)陣列模型

陣列輸出模型式(1)可修正為

陣列誤差的存在使得實(shí)際噪聲子空間與理想信號(hào)子空間不完全正交,需要進(jìn)行誤差校正。









為了論文的完整性,下面首先簡(jiǎn)要介紹文獻(xiàn)[10]基于子空間的正交原理提出的經(jīng)典聯(lián)合迭代自校正算法(WF算法),該算法通過(guò)對(duì)式(13)的優(yōu)化問(wèn)題進(jìn)行迭代求解來(lái)實(shí)現(xiàn)陣列誤差的自校正。


WF算法通常采用譜峰搜索方法來(lái)獲得信源方位估計(jì),對(duì)于共形陣列,進(jìn)行兩維譜峰搜索的運(yùn)算量大,難以實(shí)時(shí)快速完成,從而限制了WF算法在共形陣列上的應(yīng)用。為解決WF算法在共形陣列上運(yùn)算量大的問(wèn)題,同時(shí)考慮到誤差校正的復(fù)雜性,本文利用虛擬陣列進(jìn)行快速DOA估計(jì),同時(shí)結(jié)合原陣列進(jìn)行誤差估計(jì),具體流程如下:




下面以圖2所示的陣列模型為例進(jìn)行分析,原共形陣列采用圖2(a)所示的半球陣列結(jié)構(gòu),利用圖2(b)所示結(jié)構(gòu)的虛擬十字陣列對(duì)其進(jìn)行內(nèi)插變換。


圖2 陣列模型


實(shí)驗(yàn)1 半球共形陣列含有0.1 dB的幅度誤差和5°的相位誤差,信噪比為10 dB,得到不同快拍數(shù)下WF算法和本文算法在線實(shí)現(xiàn)幅相誤差校正所需的CPU時(shí)間,如表1所示。
表1在線實(shí)現(xiàn)幅相誤差校正的CPU時(shí)間對(duì)比(s)

快拍數(shù)WF算法本文算法 100224.44430.3702 200224.20720.3584 500225.91540.3723
可以看出,相比于WF算法,本文算法在線完成1次幅相誤差校正所用時(shí)間縮短了兩個(gè)數(shù)量級(jí),證實(shí)了本文算法運(yùn)算量低、實(shí)現(xiàn)快速的優(yōu)點(diǎn)。
實(shí)驗(yàn)2 半球共形陣列含有0.1 dB的幅度誤差和5°的相位誤差,快拍數(shù)為500,當(dāng)信噪比由0 dB變化至20 dB時(shí),經(jīng)過(guò)1000次獨(dú)立的蒙特卡洛實(shí)驗(yàn),得到WF算法和本文算法進(jìn)行DOA估計(jì)以及幅相誤差校正的性能對(duì)比曲線,如圖3所示。
由仿真結(jié)果可知,兩種算法的幅相誤差校正性能受信噪比影響較大,信噪比越大兩種算法的校正性能越優(yōu)。在幅度誤差校正方面,兩種算法性能接近;相位誤差校正方面,信噪比較低時(shí),本文算法性能明顯低于原算法,隨著信噪比的增大,二者校正性能接近。
實(shí)驗(yàn)3 信噪比取10 dB,相位誤差由0°變化至8°,其它條件同實(shí)驗(yàn)2,得到WF算法和本文算法進(jìn)行DOA估計(jì)和幅相誤差校正的性能對(duì)比曲線,如圖4所示。
由仿真結(jié)果可知,兩種算法對(duì)幅度誤差的校正性能受相位誤差影響較小,WF算法性能略優(yōu)于本文算法。在DOA估計(jì)和相位誤差校正方面,兩種算法的的校正性能均隨相位誤差的增大而變差,相位誤差較小時(shí),文中算法與WF算法性能接近,隨著相位誤差的增大,文中算法性能逐漸變差。簡(jiǎn)言之,文中算法對(duì)幅度誤差的校正性能受相位誤差影響較小,而在DOA估計(jì)和相位誤差校正方面對(duì)相位誤差更敏感。
實(shí)驗(yàn)4 相位誤差取5°,幅度誤差由0 dB增至0.5 dB,其它條件同實(shí)驗(yàn)3,得到兩種算法進(jìn)行DOA估計(jì)和幅相誤差校正的性能對(duì)比曲線,如圖5所示。
由仿真結(jié)果可知,在DOA估計(jì)和幅相誤差校正方面,文中算法略差于WF算法,其性能不隨幅度誤差的增大而變化,即文中算法DOA估計(jì)和幅相誤差校正性能獨(dú)立于幅度誤差。

圖3 均方根誤差隨信噪比變化對(duì)比曲線

圖4均方根誤差隨相位誤差變化對(duì)比曲線
圖5均方根誤差隨幅度誤差變化對(duì)比曲線
本文將虛擬內(nèi)插變換引入到共形陣列的幅相誤差校正中,利用虛擬陣列快速高效地實(shí)現(xiàn)對(duì)信源方位的DOA估計(jì),并利用DOA估計(jì)值得到原陣列的誤差參數(shù)。所提算法具有運(yùn)算復(fù)雜度低、實(shí)現(xiàn)快速的優(yōu)點(diǎn),可為共形陣列工程應(yīng)用提供參考。本文方法僅適合對(duì)幅相誤差進(jìn)行校正,當(dāng)陣列存在方位依賴誤差時(shí)如何進(jìn)行誤差校正,還需進(jìn)一步探討和研究。
[1] Mailloux R J. Conformal array antenna theory and design[J]., 2007, 49(5): 126-127.
[2] Belloni F, Richter A, and Koivunen V. DoA estimation via manifold separation for arbitrary array structures[J]., 2007, 55(10): 4800-4810.
[3] Blomberg A E A, Austeng A, and Hansen R E. Adaptive beamforming applied to a cylindrical sonar array using an interpolated array transformation[J]., 2012, 37(1): 25-34.
[4] Zou L, Lasenby J, and He Z. Direction and polarisation estimation using polarised cylindrical conformal arrays[J]., 2012, 6(5): 395-403.
[5] 王布宏, 郭英, 王永良, 等. 共形天線陣列流形的建模方法[J]. 電子學(xué)報(bào), 2009, 37(3): 481-484.
Wang Bu-hong, Guo Ying, Wang Yong-liang,.. Array manifold modeling for conformal array antenna[J]., 2009, 37(3): 481-484.
[6] Costa M, Richter A, and Koivunen V. DoA and polarization estimation for arbitrary array configurations[J]., 2012, 60(5): 2330-2343.
[7] Boon C N and See C S. Sensor-array calibration using a maximum-likelihood approach[J]., 1996, 44(6): 827-835.
[8] See C M S. Sensor array calibration in the presence of mutualcoupling and unknown sensor gains and phases[J]., 1994, 30(5): 373-374.
[9] See C M S. Method for array calibration in high-resolution sensor array processing[J].,, 1995, 142(3): 90-96.
[10] Friedlander B and Weiss A J. Eigenstructure methods for direction finding with sensor gain and phase uncertainties[C]. Proceedings of the International Conference on Acoustics, Speech, and Signal Processing, New York, 1988: 2681-2684.
[11] Liu Ai-fei, Liao Gui-sheng, Zeng Cao,.. An eigenstructure method for estimating DOA and sensor gain-phase errors[J]., 2011, 59(12): 5944-5956.
[12] Cao Sheng-hong, Ye Zhong-fu,.. DOA estimation based on fourth-order cumulants in the presence of sensor gain-phase errors[J]., 2013, 93(9): 2581-2585.
[13] Cao Sheng-hong, Ye Zhong-fu, Xu Dong-yang,.. A Hadamard product based method for DOA estimation and gain-phase error calibration[J]., 2013, 49(2): 1224-1233.
[14] Friedlander B. The root-MUSIC algorithm for direction finding with interpolated arrays[J]., 1993, 30(1): 15-29.
[15] Weiss A J and Gavish M. Direction finding using ESPRIT with interpolated arrays[J]., 1991, 39(6): 1473-1478.
張學(xué)敬: 男,1988年生,碩士生,研究方向?yàn)楣残侮嚵行盘?hào)處理.
楊志偉: 男,1980年生,副教授,研究方向?yàn)殛嚵行盘?hào)處理、地面運(yùn)動(dòng)目標(biāo)檢測(cè)、極化處理.
廖桂生: 男,1963年生,教授,研究方向?yàn)樽赃m應(yīng)信號(hào)處理、信號(hào)檢測(cè)與估計(jì).
A Fast Method for Gain-phase Error Calibration in Conformal Array
Zhang Xue-jing Yang Zhi-wei Liao Gui-sheng
(,,710071,)
The joint estimation of the Direction of Arrival (DOA) and gain-phase errors can be implemented by the joint iteration method based on the eigen structure subspaces. However, when applying the method to correct the amplitude and phase error of the conformal array, the fast high-resolution spatial spectrum estimation methods can not be applied directly, because of that the space-domain steering vectors of the conformal array does not possess the Vander monde structure. On the other side, the computation of DOA estimation implemented by searching peak of spatial spectrum in 2-dimension is very large, which limits the application of joint iteration method in conformal array. To solve this problem, this paper proposes a new method for gain-phase error calibration in conformal array by virtual interpolation. The DOA estimation can be implemented rapidly by utilizing the special structure of virtual array, and the searching process of the spatial spectrum peak is eliminated, thus the computational complexity of the proposed method is low and the engineering realization of the proposed method is easy. Theoretical analysis and extensive simulations verify the effectiveness of the proposed methods, and provide a reference for the engineering applications of conformal arrays.
DOA estimation; Gain-phase error calibration; Conformal array; Virtual interpolation
TN911.7
A
1009-5896(2014)05-1100-06
10.3724/SP.J.1146.2013.01025
張學(xué)敬xjzhang7@163.com
2013-07-11收到,2013-11-07改回
長(zhǎng)江學(xué)者和創(chuàng)新團(tuán)隊(duì)發(fā)展計(jì)劃(IRT-0954),國(guó)家自然科學(xué)基金(60901066),中央高校基本科研業(yè)務(wù)費(fèi)專(zhuān)項(xiàng)資金(K5051302007)和陜西省教育廳科研計(jì)劃項(xiàng)目(2013JK1051)資助課題