Page 204 - 201805

786 2018 年 9 月

由图7(a)可知，MVDR算法无法实现这两目标 [9] Cadzow J A. A high resolution direction-of-arrival algo-
的分辨；由图7(b)可知，GMUSIC算法与MUSIC算 rithm for narrow-band coherent and incoherent sources[J].
IEEE Transactions on Acoustics Speech & Signal Process-
法都可将这两个目标正确分辨出，但GMUSIC算法
ing, 1988, 36(7): 965–979.
的性能要好于MUSIC算法。 [10] Viberg M, Ottersten B, Kailath T. Detection and estima-
tion in sensor arrays using weighted subspace fitting[J].
5 结论 IEEE Transactions on Signal Processing, 1991, 39(11):
2436–2449.
本 文 深 入 研 究 了 MUSIC 算 法 的 改 进 算 法 [11] Clergeot H, Tressens S, Ouamri A. Performance of high
GMUSIC，并与 MUSIC 算法的性能进行了仿真 resolution frequencies estimation methods compared to
the Cramer-Rao bounds[J]. IEEE Transactions on Acous-
分析与水池试验对比。GMUSIC算法是根据随机矩 tics Speech & Signal Processing, 1989, 37(11): 1703–1720.
阵理论与 G 估计理论而提出，由上述仿真分析与水 [12] Bilik I. Spatial compressive sensing for direction-of-arrival
池试验可知，相较于 MUSIC算法，GMUSIC算法有 estimation of multiple sources using dynamic sensor ar-
rays[J]. IEEE Transactions on Aerospace and Electronic
以下几个方面的优点：(1)GMUSIC 算法可以更好
Systems, 2011, 47(3): 1754–1769.
地分辨相邻目标，且需要的快拍数较 MUSIC 算法 [13] Kim J M, Lee O K, Ye J C. Compressive MUSIC: revisit-
要少；(2) 在低信噪比情况下，GMUSIC 算法方位估 ing the link between compressive sensing and array signal
计均方根误差远小于 MUSIC 算法，估计成功概率 processing[J]. IEEE Transactions on Information Theory,
2012, 58(1): 278–301.
远大于 MUSIC 算法。上述仿真分析与水池试验验
[14] Wang Y, Geert L, Ashish P. Direction estimation using
证还表明，GMUSIC 方法在水声信号处理中也确实 compressive sampling array processing[C]. United King-
可以实现小快拍方位估计，但对快拍数的大小变化 dom Wales Cardiff: Statistical Signal Processing, 2009.
较为敏感，在水下小尺度运动阵列的小快拍方位估 SSP’09. IEEE/SP 15th Workshop, 2009: 626–629.
[15] Gorodnitsky I F, Rao B D. Sparse signal reconstruction
计中有一定的应用价值。
from limited data using FOCUSS: a re-weighted minimum
norm algorithm[J]. IEEE Transactions on Signal Process-
ing, 1997, 45(3): 600–616.
参 考 文 献
[16] Cotter S F, Rao B D, Engan K, et al. Sparse solutions to
linear inverse problems with multiple measurement vec-
[1] Krim H, Viberg M. Two decades of array signal processing
tors[J]. IEEE Transactions on Signal Processing, 2005,
research: the parametric approach[J]. IEEE Signal Pro-
53(7): 2477–2488.
cessing Magazine, 1996, 13(4): 67–94.
[17] Malioutov D, Çetin M, Willsky A S. A sparse signal recon-
[2] Kay S M, Marple S L J. Spectrum analysis—A mod-
struction perspective for source localization with sensor
ern perspective[J]. Proceedings of the IEEE, 1981, 69(11):
arrays[J]. IEEE Transactions on Signal Processing, 2005,
1380–1419.
53(8): 3010–3022.
[3] Burg J P. Maximum entropy spectral analysis[C]. Okla-
[18] Mestre X, Lagunas M Á. Modified subspace algorithms for
homa City: 37th Ann. Int. Meeting Soc. Explor. Geo-
DOA estimation with large arrays[J]. IEEE Transactions
phys, 1967: 141–145.
on Signal Processing, 2008, 56(2): 598–614.
[4] Capon J. High-resolution frequency-wavenumber spec-
trum analysis[J]. Proceedings of the IEEE, 1969, 57(8): [19] Mestre X. Improved estimation of eigenvalues and eigen-
1408–1418. vectors of covariance matrices using their sample es-
[5] Schmidt R O. Multiple emitter location and signal param- timates[J]. IEEE Transactions on Information Theory,
2008, 54(11): 5113–5129.
eter estimation[J]. IEEE Transactions on Antennas and
Propagation, 1986, 34(3): 276–280. [20] 曾杏元. 有关随机矩阵领域最新研究动态与进展的综述报
[6] Roy R, Paulraj A, Kailath T. ESPRIT—A subspace ro- 告 [J]. 数学理论与应用, 2011, 31(3): 7–19.
tation approach to estimation of parameters of cisoids in Zeng Xingyuan. A survey on the latest dynamic and
noise[J]. IEEE Transactions on Acoustics Speech & Signal progress of random matrix field[J]. Mathematical Theory
Processing, 1986, 34(5): 1340–1342. and Applications, 2011, 31(3): 7–19.
[7] Stoica P, Nehorai A. MUSIC, maximum likelihood, and [21] 金百锁. 大维随机矩阵谱分布的极限理论研究及其应用 [D].
Cramer-Rao bound[J]. IEEE Transactions on Acoustics 合肥: 中国科学技术大学, 2006: 12–30.
Speech & Signal Processing, 1989, 37(5): 720–741. [22] Girko V. An introduction to statistical analysis of random
[8] Ottersten B, Viberg M, Stoica P, et al. Exact and large arrays[M]. The Netherlands: VSP, 1998: 40–45.
sample maximum likelihood techniques for parameter es- [23] Bai Z D, Miao B Q, Pan G M. On asymptotics of eigen-
timation and detection in array processing[M]. Berlin Hei- vectors of large sample covariance matrix[J]. The Annals
delberg: Springer Verlag, 1993: 99–151. of Probability, 2007, 35(4): 1532–1572.