单矢量水听器的改进稀疏贝叶斯学习方位估计算法  

Improved sparse Bayesian learning direction estimation algorithm for single vector hydrophones

作　　者：李雄辉 梁国龙[1,3,4] 沈同圣 罗再磊 LI Xionghui;LIANG Guolong;SHEN Tongsheng;LUO Zailei(College of Underwater Acoustic Engineering,Harbin Engineering University,Harbin 150001;National Innovation Institute of Defense Technology,Advanced Interdisciplinary Technology Research Center,Beijing 100071;Key Laboratory of Marine Information Acquisition and Security(Harbin Engineering University),Ministry of Industry and Information Technology,Harbin 150001;National Key Laboratory of Underwater Acoustic Technology,Harbin Engineering University,Harbin 150001)

机构地区：[1]哈尔滨工程大学水声工程学院,哈尔滨150001 [2]国防科技创新研究院前沿交叉技术研究中心,北京100071 [3]工业和信息化部海洋信息获取与安全工信部重点实验室(哈尔滨工程大学),哈尔滨150001 [4]哈尔滨工程大学水声技术重点实验室,哈尔滨150001

出　　处：《声学学报》2025年第1期77-85,共9页Acta Acustica

基　　金：国家自然科学基金项目(41906169)资助。

摘　　要：复数稀疏贝叶斯学习(SBL)算法计算量大,为此将声能流与稀疏贝叶斯学习算法相结合,提出了基于单矢量水听器的声能流稀疏贝叶斯学习(SI-SBL)方位估计算法。该方法采用声能流取代声压振速信息作为观测量,将参数估计过程从复数域运算转化为实数域运算,同时利用声压通道噪声与振速噪声不相关的特点实现了噪声抑制,进一步加快了稀疏贝叶斯学习算法收敛速度,使SI-SBL算法获得相比以声压振速通道作为观测量的SBL算法更高的估计精度和尖锐的谱峰。仿真数据表明,单矢量水听器SI-SBL算法相比于SBL算法具有更高的精度和更快的计算速度。实验数据验证,SI-SBL算法相比SBL精度提高了25%,运算速度提高了8倍,证明了本文所提SI-SBL算法应用于水平方位估计的可行性。Aiming at the problem that complex domain sparse Bayesian learning(SBL)requires a lot of computation,a sound intensity based sparse Bayesian learning(SI-SBL)method,which combine sound intensity with SBL algorithms,is proposed to estimate the source direction-of-arrivals by single vector hydrophone.The SI-SBL algorithm converts the parameter estimation process from complex domain operations to real number domain operations by using the sound intensities as an observation.Meanwhile,noise suppression is realized by using the feature that the sound pressure channel is not related to the vibration velocity noise,which further accelerates the rate of convergence of the sparse Bayesian learning algorithm,and enables the SI-SBL algorithm to obtain higher estimation accuracy and sharp spectral peaks than SBL algorithm.The simulation results show that the single vector hydrophone SI-SBL algorithm not only performs better than the SBL algorithm,but also reduces computational complexity.The experimental results show that the SI-SBL algorithm has a 25%improvement in accuracy and an 8-fold increase in computational speed compared to SBL algorithm,verifying the effectiveness of SI-SBL algorithm.

关 键 词：矢量水听器 方位估计 稀疏贝叶斯学习 声能流 

分 类 号：TB565.1[交通运输工程—水声工程] TP18[理学—物理]