方慶園 韓 勇 金 銘 宋立眾② 喬曉林
?
基于噪聲子空間特征值重構(gòu)的DOA估計(jì)算法
方慶園①韓 勇*①金 銘①宋立眾①②喬曉林①
①(哈爾濱工業(yè)大學(xué)電子與信息工程學(xué)院 哈爾濱 150001)②(毫米波國(guó)家重點(diǎn)實(shí)驗(yàn)室 南京 210096)
該文針對(duì)非等功率信號(hào)波達(dá)方向(DOA)估計(jì)問(wèn)題,提出一種基于噪聲子空間特征值重構(gòu)(Eigenvalue Reconstruction of Noise Subspace, ERNS)的超分辨算法。算法對(duì)接收信號(hào)自相關(guān)矩陣進(jìn)行特征值分解,通過(guò)重構(gòu)噪聲空間特征值以及引入虛擬信源來(lái)構(gòu)造新的接收信號(hào)自相關(guān)矩陣,對(duì)該矩陣進(jìn)行特征值分解得到新的噪聲空間特征值。當(dāng)虛擬信源與實(shí)際信源入射方向相同時(shí),新噪聲空間特征值與重構(gòu)后噪聲空間特征值保持不變,利用這一特性來(lái)估計(jì)信源入射方向。該文給出算法的原理及實(shí)現(xiàn)步驟,并通過(guò)仿真進(jìn)行原理驗(yàn)證與性能分析,仿真結(jié)果表明與其他子空間算法和MUSIC 算法相比,ERNS算法能夠提高弱信號(hào)估計(jì)成功的概率。
陣列信號(hào)處理;高分辨率;波達(dá)方向估計(jì);噪聲子空間;特征值重構(gòu)
當(dāng)存在相鄰非等功率信號(hào)在同一波束內(nèi)入射時(shí),強(qiáng)信號(hào)對(duì)弱信號(hào)的壓制會(huì)使得MUSIC等超分辨估計(jì)算法對(duì)弱信號(hào)的估計(jì)發(fā)生較大偏差甚至無(wú)法估計(jì)[8,9]。針對(duì)不等功率入射信號(hào)的DOA估計(jì)目前大致有兩種解決方案:(1)采用極大似然估計(jì)類(lèi)算法同時(shí)估計(jì)出強(qiáng)弱信號(hào)的波達(dá)角[10,11],(2)在強(qiáng)信號(hào)波達(dá)方向已知時(shí),通過(guò)干擾阻塞方法對(duì)強(qiáng)信號(hào)進(jìn)行抑制來(lái)估計(jì)弱信號(hào)的DOA[12,13]。若強(qiáng)信號(hào)入射方向未知,則需首先確定強(qiáng)信號(hào)入射方向。例如文獻(xiàn)[12]使用干擾阻塞法,利用強(qiáng)干擾信號(hào)的先驗(yàn)知識(shí)構(gòu)造阻塞矩陣來(lái)抑制已知方位的強(qiáng)干擾,實(shí)現(xiàn)了特定區(qū)域內(nèi)的低信噪比信號(hào)方位估計(jì)。但該算法僅限于均勻線陣且需預(yù)知強(qiáng)干擾的入射方向。對(duì)阻塞類(lèi)算法來(lái)說(shuō)當(dāng)強(qiáng)弱信號(hào)角度相近時(shí),算法在抑制強(qiáng)信號(hào)的同時(shí)也衰減了弱信號(hào)[14]。文獻(xiàn)[13]將強(qiáng)信號(hào)導(dǎo)向矢量所在空間納入噪聲子空間,再在該擴(kuò)展的噪聲子空間上利用常規(guī)MUSIC算法進(jìn)行弱信號(hào)的DOA估計(jì)。當(dāng)強(qiáng)弱信號(hào)夾角較小時(shí)該方法也存在抑制強(qiáng)信號(hào)的同時(shí)衰減弱信號(hào)的問(wèn)題[15],只是對(duì)強(qiáng)信號(hào)的抑制較大??梢?jiàn)阻塞類(lèi)算法與擴(kuò)展噪聲類(lèi)算法均存在強(qiáng)信號(hào)對(duì)弱信號(hào)的壓制問(wèn)題。與上述兩種方案不同,文獻(xiàn)[16]提出了基于噪聲子空間特征值不變性的超分辨估計(jì)算法,該方法在入射信號(hào)功率相等或不等條件下均有較好的估計(jì)性能。
在文獻(xiàn)[16]基礎(chǔ)上,本文提出一種噪聲空間特征值重構(gòu)的DOA估計(jì)方法。通過(guò)優(yōu)化噪聲空間特征值的分布,使得噪聲空間特征值改變量對(duì)虛擬輻射源A的入射角度更加敏感從而提高了算法的估計(jì)性能。仿真實(shí)驗(yàn)證明了本文所提出的基于噪聲子空間特征值重構(gòu)(ERNS)DOA估計(jì)方法的有效性。
式中為陣列接收數(shù)據(jù)矢量,T表示矩陣轉(zhuǎn)置;為維噪聲矢量;為維信號(hào)矢量;為維陣列導(dǎo)向矢量。
其中
由定理1構(gòu)造算法的空間譜表達(dá)式為
將式(20)代入式(19)得
因此,由式(22)可得
整理式(30)得
根據(jù)第3節(jié)證明,本算法DOA估計(jì)步驟如下。
本文提出了一種基于噪聲空間特征值重構(gòu)的DOA估計(jì)算法。通過(guò)重構(gòu)噪聲空間特征值,減小了信號(hào)空間泄漏到噪聲空間的能量,從而提高了算法性能。本文所提算法對(duì)天線接收陣列形式無(wú)特殊要求,對(duì)1維與2維DOA估計(jì)均適用。
圖2 隨兩信號(hào)方位夾角變化弱信號(hào)估計(jì)成功的概率
圖3 不同信噪比下弱信號(hào)估計(jì)的成功概率
圖4 隨弱信號(hào)信噪比變化弱信號(hào)估計(jì)值的均方根誤差
[1] Agarwal K, Li Pan, Legong Y K,.. Parctical applications of multiple signal classification[J].-, 2012, 22(3): 359-369.
[2] Rangarao K V and Venkatanarasimhan S. Gold-MUSIC: a variation on MUSIC to accurately determine peaks of the spectrum[J].2013, 61(4): 2263-2268.
[3] 閆鋒剛, 金銘, 喬曉林. 適用任意陣列的變換域二維波達(dá)角快速估計(jì)算法[J]. 電子學(xué)報(bào), 2013, 41(5): 936-942.
Yan Feng-gang, Jin Ming, and Qiao Xiao-lin. Fast 2-D DOA estimation method in transformed domain with arbitrary arrays[J]., 2013, 41(5): 936-942.
[4] 王凌, 李國(guó)林, 謝鑫, 等. 非圓信號(hào)二維DOA和初始相位聯(lián)合估計(jì)方法[J]. 雷達(dá)學(xué)報(bào), 2012, 1(1): 43-49.
Wang Ling, Li Guo-lin, Xie Xin,.. Joint 2-D DOA and noncircularity phase estimation method[J]., 2012, 1(1): 43-49.
[5] Qu J Y, Li X, and Wen Y J. A new method for weak signals’ DOA estimation in the presence of strong interferences[C]. International Conference on Signal Processing, Beijing, 2012: 320-323.
[6] He J, Swamy M N S, and Ahmad M O. Joint DOD and DOA estimation for MIMO array with velocity receive sensors[J]., 2011, 18(7): 399-402.
[7] 朱偉, 陳伯孝. 強(qiáng)相干干擾下微弱信號(hào)波達(dá)方向估計(jì)[J]. 電波科學(xué)學(xué)報(bào), 2013, 28(2): 212-219.
Zhu Wei and Chen Bai-xiao. Weak signal DOA estimation under coherent intensive interferences[J]., 2013, 28(2): 212-219.
[8] 柴立功, 羅景青. 一種強(qiáng)干擾條件下微弱信號(hào)DOA估計(jì)的新方法[J]. 電子與信息學(xué)報(bào), 2005, 27(10): 1517-1520.
Chai Li-gong and Luo Jing-qing. A novel algorithm for weak signals’ DOA estimation under intensive interferences[J].&, 2005, 27(10): 1517-1520.
[9] 程正東, 羅景青, 樊祥, 等. 信號(hào)源功率不一致對(duì)MUSIC算法分辨性能的影響[J]. 電子與信息學(xué)報(bào), 2008, 30(5): 1088-1091.
Cheng Zheng-dong, Luo Jing-qing, Fan Xiang,.. Effect of power difference of two signal sources on resolving performance of MUSIC algorithm[J].&, 2008, 30(5): 1088-1091.
[10] Stoica P and Sharman K C. Maximum likelihood methods for direction-of-arrival estimation[J]., 1990, 38(7): 1132-1143.
[11] Cedervall M and Moses R L. Efficient maximum likelihood DOA estimation for signals with known waveforms in the presence of multipath[J]., 1997, 45(3): 808-812.
[12] 陳輝, 蘇海軍. 強(qiáng)干擾/信號(hào)背景下的 DOA估計(jì)新方法[J]. 電子學(xué)報(bào). 2006, 34(3): 530-534.
Chen Hui and Su Hai-jun. A new approach to estimate DOA in presence of strong jamming/signal suppression[J]., 2006, 34(3): 530-534.
[13] 張靜, 廖桂生, 張潔. 強(qiáng)信號(hào)背景下基于噪聲子空間擴(kuò)充的弱信號(hào)DOA估計(jì)方法[J]. 系統(tǒng)工程與電子技術(shù), 2009, 31(6): 1279-1283.
Zhang Jing, Liao Gui-sheng, and Zhang Jie. DOA estimation based on extended noise subspace in the presence of strong signals[J]., 2009, 31(6): 1279-1283.
[14] 賀順, 楊志偉, 張娟, 等. 自適應(yīng)加權(quán)修正的強(qiáng)弱信號(hào)Capon譜估計(jì)方法[J]. 系統(tǒng)工程與電子技術(shù), 2013, 35(5): 905-908.
He Shun, Yang Zhi-wei, Zhang Juan,.. Modified Capon approach with adaptive weighted for discriminating strong and weak signals[J].2013, 35(5): 905-908.
[15] 徐亮, 曾操, 廖桂生, 等. 基于特征波束形成的強(qiáng)弱信號(hào)波達(dá)方向與新源數(shù)估計(jì)方法[J]. 電子與信息學(xué)報(bào). 2011, 33(2): 321-325.
Xu Liang, Zeng Cao, Liao Gui-sheng,.. DOA and source number estimation method for strong and weak signals based on eigen beamforming[J].&, 2011, 33(2): 321-325.
[16] Olfat A and Nader-Esfahani S. A new signal subspace processing for DOA estimation[J]., 2004, 84(4): 721-728.
方慶園: 女,1987年生,博士生,研究方向?yàn)殛嚵行盘?hào)處理與天線技術(shù).
韓 勇: 男,1976年生,講師,主要研究方向?yàn)殛嚵行盘?hào)處理、DOA 估計(jì)及目標(biāo)識(shí)別.
DOA Estimation Based on Eigenvalue Reconstruction of Noise Subspace
Fang Qing-yuan①Han Yong①Jin Ming①Song Li-zhong①②Qiao Xiao-lin①
①(,,150001,)②(,210096,)
This paper proposes an Eigenvalue Reconstruction method in Noise Subspace (ERNS) for Direction of Arrival DOA estimation with high resolution, provided that the powers of sources are different. The noise subspace eigenvalues belonging to the covariance matrix of received signals, obtained by EigenValue Decomposition (EVD), are modified to construct a new covariance matrix with respect to virtual source. The noise subspace eigenvalues corresponding to the new covariance matrix remain the same as before they are modified. The invariance of the noise subspace is utilized to estimate the DOA of emitters. The theory and process of ERNS algorithm are provided, at the same time, the theory and performance of ERNS algorithm is validated by computer simulations. The simulation results show that the ERNS algorithm has a better performance in successful probability of weak signal estimation compared with other subspace methods and MUSIC algorithm.
Array signal processing; High resolution; Direction of Arrival (DOA) estimation; Noise subspace; Eigenvalue reconstruction
TN911.7
A
1009-5896(2014)12-2876-06
10.3724/SP.J.1146.2013.02014
韓勇 han8662033@163.com
2013-12-23收到,2014-04-15改回
國(guó)家自然科學(xué)基金(61171181),毫米波國(guó)家重點(diǎn)實(shí)驗(yàn)室開(kāi)放課題(K201328),哈爾濱工業(yè)大學(xué)科技創(chuàng)新基金(HIT.NSRIF2013130)和哈爾濱工業(yè)大學(xué)(威海)校科學(xué)研究基金(HIT(WH)XBQD201022)資助課題