一种快速稀疏贝叶斯学习的水声目标方位估计方法研究  

A fast underwater acoustic target direction of arrival estimation method based on sparse Bayesian learning

作　　者：王彪[1,2] 朱志慧[1] 戴跃伟[1] 

机构地区：[1]江苏科技大学电子信息学院,镇江212003 [2]东南大学水声信号处理教育部重点实验室,南京210000

出　　处：《声学学报》2016年第1期81-86,共6页Acta Acustica

基　　金：国家自然科学基金(11204109;11574120);江苏省高校自然科学基金(12KJB510003;13KJB510007);东南大学水声信号处理教育部重点实验室开放基金(UASP1503);江苏省青蓝工程;江苏高校优势学科建设工程-船舶与海洋工程及江苏科技大学青年学者计划项目资助

摘　　要：针对以具有时序结构的稀疏贝叶斯学习(Temporally multiple sparse Bayesian learning,TMSBL)为重构算法的水声目标DOA(Direction-of-arrival)估计方法存在运算速度慢的问题,结合块稀疏贝叶斯学习(Block-spare Bayesian learning,BSBL)理论框架下DOA估计模型与特点,采用MacKay提出的定点方法(Fixed-point method)对TMSBL算法中的核心超参量进行求解,提出一种快速的水声目标方位估计稀疏贝叶斯学习的方法,该方法具有运算速度快,重构概率高的特点,并通过实验仿真从运算时间、失败率和均方根误差等方面与TMSBL算法进行比较,验证了该方法的可行性与有效性。The direction of arrival （DOA） estimation methods for underwater acoustic target which using temporally multiple sparse Bayesian learning （TMSBL） as the reconstructing algorithm, were slower for the process of computation. To solve this problem, a fast underwater acoustic target direction of arrival estimation based on sparse Bayesian learning research was proposed. Analyzing the model characteristics of block-spare Bayesian learning framework for DOA, the algorithm proposed to obtain the value of core hyper-parameter through MacKay＇s fixed-point method to estimate the DOA, so it spent less time for computation, and had more superior recovery performance than TMSBL algorithm. Simulation results verified the feasibility and effectiveness of the proposed algorithm by comparing the DOA estimation performance with TMSBL and other traditional DOA algorithms through different facets such as operation time, failure rate and root mean square error.

关 键 词：贝叶斯学习 目标方位估计 水声目标 估计方法 稀疏 LEARNING 重构算法 运算速度 

分 类 号：O427[理学—声学] TP18[理学—物理]