西北工业大学航海学院,西安,710072
网络首发:2009-12-10,
纸质出版:2009
移动端阅览
侯云山, 黄建国, 金勇. 被动合成阵列极大似然参数估计的快速方法[J]. 西安交通大学学报, 2009,43(12):106-110.
Fast Algorithm of the Maximum Likelihood Parameter Estimation for Passive Synthetic Arrays[J]. 2009, 43(12): 106-110.
针对被动合成阵列的极大似然目标参数估计方法(pasaML)计算量大的问题
探索利用Markov Monte Carlo(MCMC)类方法降低计算量.将完备抽样方法(PS)与pasaML方法相结合
提出一种基于完备抽样的频率-方位联合估计新方法(PS-pasaML)来联合估计多个目标的频率和方位.首先将pasaML方法的谱函数视为频率和方位的联合概率密度函数
构造并证明了具有单调保偏序性质的更新函数
然后产生2条Markov链的初始方位角向量
使用更新函数确定Markov链在状态空间中的转移方向
并通过在全局状态空间和局部状态空间之间的跳转抽样来提高运算速度
最后由融合时间判决Markov链的平稳性
对其求期望从而获得目标方位的最终估计.仿真实验表明
在目标个数较少时
PS-pasaML方法不仅保持了pasaML方法的高分辨能力
而且计算复杂度降低为pasaML方法的1/7左右.
A class of Markov Monte Carlo methods is explored to reduce the heavy computation load of the maximum likelihood parameter estimation method for passive synthetic arrays(pasaML). The method combines the perfect sampling technique with the pasaML method to form a frequency-azimuth joint estimation method(called PS-pasaML)to estimate the frequencies and directions of multiple sources at the same time
The power of the pasaML spectrum function is regarded as the target distribution of azimuths and frequency
and an updating function with monotone property is constructed and proved. Then the initial azimuth vectors of two Markov chains are produced and the updating function is used to determine the transfer directions of the Markov chains in the state space. The computational speed is accelerated by switching the sampling process between the global state space and local state spaces. The stationary of the Markov process is judged by the coalescence time
and the expectation of the stationary yields the final estimation of the target azimuths. Simulation results show that when the source number is small
the proposed method
retains the high-resolution performance of the pasaML method
and the computational complexity is only 1/7 of the pasaML method.
侯云山,黄建国,张立杰,等.一种新的浅海目标方位估计方法[J]. 西安交通大学学报, 2008, 42(10): 1295-1299.
HOU Yunshan, HUANG Jianguo, ZHANG Lijie,et al. A novel bearing estimation method for shallow water targets [J]. Journal of Xi'an Jiaotong University, 2008, 42(10): 1295-1299.
李宇,黄勇,黄海宁. 空时被动合成孔径阵列处理算法研究[J]. 信号处理, 2008, 24(3): 426-430.
LI Yu, HUANG Yong, HUANG Haining. Spatial-temporal processing methods for passive synthetic aperture array [J]. Signal Processing, 2008, 24(3): 426-430.
SULLIVAN E J, HOLMES J D, CAREY W M, et al. Broadband passive synthetic aperture: experimental results [J]. J Acoust Soc Am, 2006, 120(4): 71-77.
STERGIOPOULOS S, URBAN H. A new passive synthetic aperture technique for towed arrays [J]. IEEE Journal of Oceanic Engineering, 1992, 17(1): 16-25.
LAROCQUE J R, REILLY J P. Reversible jump MCMC for joint detection and estimation of sources in colored noise [J]. IEEE Transactions on Signal Processing, 2002, 52(2): 231-240.
DJURIC P M, HUANG Yufei,TADESSE G. Perfect sampling: a review and applications to signal processing [J]. IEEE Transactions on Signal Processing, 2002, 50(2): 345-356.
GUO Qinghua, LIAO Guisheng. Fast ML spectrum peak search via Metropolis-Hastings sample [J]. Journal of Electronics(China), 2005, 22(6): 599-604.
HUANG Jianguo, SUN Yi, LIU Kewei, et al. A new Gibbs sampling DOA estimator based on Bayesian method [C]∥Proceedings of ICASSP.Piscataway,NJ, USA: IEEE, 2003: 229-232.
金勇,黄建国,张立杰. 短采样宽带信号方位估计快速算法[J].电子与信息学报, 2008, 30(10): 2525-2528.
JIN Yong, HUANG Jianguo, ZHANG Lijie. DOA estimation fast algorithm for short sampling wideband signals [J]. Journal of Electronics Information Technology, 2008, 30(10): 2525-2528.
0
浏览量
4
下载量
2
CSCD
关联资源
相关文章
相关作者
相关机构
京公网安备11010802024621