Method for Estimating the Location of A Low-frequency Target in A Shallow Sea Based on A Single Vector Hydrophone

To estimate the location of a low-frequency target in the complex environment of shallow seas, both real-time properties and accuracy must be considered. In this paper, a multichannel information fusion method is proposed for estimating the location of a target based on a single vector hydrophone. First, a window signal fusion algorithm is proposed, which combines the EM algorithm to achieve adaptive signal extraction. The RNN used later realized the self-localization of the sound source. The results show that the fixed-dynamic window based on the steepest rising segment of Shannon entropy can divide a very short signal sample into a sufficient number of signal sets and the signal self-replenishment based on the EM algorithm can further strengthen the characteristics of different signal segments on the basis of replenishing the signal. Compared with other networks, the RNN has the highest accuracy and stability for this type of location estimation of an acoustic source based on the time-domain signal. The experimental results indicate that in the shallow sea environment, this model only depends on a single vector hydrophone to collect signals and rapidly locate the acoustic source, and the error radius is controlled within 1.5 m.

[1]  Monika Agrawal,et al.  Iterative adaptive approach to DOA estimation with acoustic vector sensors , 2015, OCEANS 2015 - Genova.

[2]  Lin Ma,et al.  An Improved Azimuth Angle Estimation Method with a Single Acoustic Vector Sensor Based on an Active Sonar Detection System , 2017, Sensors.

[3]  G. McLachlan,et al.  The EM algorithm and extensions , 1996 .

[4]  Yajie Miao,et al.  EESEN: End-to-end speech recognition using deep RNN models and WFST-based decoding , 2015, 2015 IEEE Workshop on Automatic Speech Recognition and Understanding (ASRU).

[5]  M. Taroudakis.  Ocean Acoustic Tomography , 2006 .

[6]  W. Marsden.  I and J , 2012 .

[7]  Hao Wang,et al.  Neural-network-based modelling and analysis for time series prediction of ship motion , 2017 .

[8]  Sander Bohte,et al.  Editorial: Artificial Neural Networks as Models of Neural Information Processing , 2017, Front. Comput. Neurosci..

[9]  Lukás Burget,et al.  Recurrent neural network based language model , 2010, INTERSPEECH.

[10]  Ting Liu,et al.  Recent advances in convolutional neural networks , 2015, Pattern Recognit..

[11]  Yoshua Bengio,et al.  Learning Phrase Representations using RNN Encoder–Decoder for Statistical Machine Translation , 2014, EMNLP.

[12]  Max Schuster,et al.  Propeller cavitation noise and background noise in the sea , 2016 .

[13]  Sander M. Bohte,et al.  Editorial: Artificial Neural Networks as Models of Neural Information Processing , 2017, Front. Comput. Neurosci..

[14]  Gary R. Wilson,et al.  Vector sensors and vector sensor line arrays: Comments on optimal array gain and detection , 2006 .

[15]  Ignacio Rojas,et al.  Neural networks: An overview of early research, current frameworks and new challenges , 2016, Neurocomputing.

[16]  K. Qu,et al.  Low frequency ambient noise modeling and comparison with field measurements in the South China Sea , 2019, Applied Acoustics.

[17]  Jung Sik Jeong,et al.  Extraction of Reference Seaway through Machine Learning of Ship Navigational Data and Trajectory , 2017, Int. J. Fuzzy Log. Intell. Syst..

[18]  Michael B. Porter,et al.  Computational Ocean Acoustics , 1994 .

[19]  Michael S. Hughes,et al.  Analysis of digitized waveforms using Shannon entropy , 1993 .