Underwater Acoustic Research Trends with Machine Learning: Active SONAR Applications

Abstract

Underwater acoustics is the study of the phenomena related to the generation, propagation, transmission and reception of sound waves in water. It is applied in a variety of underwater activities such as underwater communication, target detection, and investigation of marine resources and environments, mainly using sound navigation and ranging (SONAR) systems. The main objective of underwater acoustic remote sensing is to indirectly acquire information on a target of interest using acoustic data. To extract information from acoustic data, machine learning, which has been recently attracting significant attention, has been employed in a variety of ways. The machine learning techniques mainly used in underwater acoustics and their applications in passive SONAR systems are introduced in the first two parts of this work, respectively (Yang et al., 2020a; Yang et al., 2020b). In the review article, we review the research on the application of machine learning in active SONAR systems for target detection and classification. 2. Active SONAR Signal Processing The passive SONAR-based target localization technique discussed in the previous part (Yang et al., 2020b) can be applied to active SONAR systems without significant modification. However, a key difference between passive and active SONAR target detection is that, in passive SONAR systems, sounds generated by targets of interest such as ships and fish are received, whereas in active SONAR systems, the observer directly transmits a signal and receives a scattered signal from the target. Consequently, for active SONAR detection, various techniques have been developed to utilize the characteristics of the sound source or those of a scattered signal depending on the properties of the target, unlike in the case of passive SONAR detection. This review article is aimed primarily at discussing active target detection and classification. 2.1 Active Target Detection and Classification Traditional active SONAR signal processing can be largely classified into the processes of (1) detecting the signal of interest, (2)

Underwater Acoustic Research Trends with Machine Learning: Ocean Parameter Inversion Applications

“…In underwater acoustics, inversion can be largely divided into the localization of surface vehicles and underwater vehicles (Parvulescu and Clay, 1965;Clay, 1966;Clay, 1987;Clay and Li, 1988;Bucker, 1976;Baggeroer et al, 1993;Tolstoy, 1993); tomography inversion, wherein the inversion operation is performed on the physical properties of seawater, such as the temperature profile over a broad area of water (Shang, 1989;Tolstoy et al, 1991;Tolstoy, 1992); and geoacoustic inversion, which yields the composition, morphology, and geological properties of the marine sediment (Rajan et al, 1987;Lynch et al, 1991). The localization of underwater sound sources has been covered in the previous parts of this review work (Yang et al, 2020b;Yang et al, 2020c). The results and trends of tomography inversion and geoacoustic inversion are reviewed herein.…”

“…Machine learning, which has recently achieved substantial success in information acquisition and extraction, is actively utilized in remote sensing. In the previous parts of this work, the research trends in the machine learning techniques and theories that are mainly used in underwater acoustics and their applications in active and passive SONAR systems were reviewed (Yang et al, 2020a;Yang et al, 2020b;Yang et. al., 2020c).…”

Accurate Identification for CW Direct Signal in Underwater Acoustic Ranging

“…Recently, ML integrated with underwater acoustics has gained significant attention, which is a type of optimization method from the statistical perspective [18][19][20][21]. The decision tree (DT) technique is commonly used and efficient for classification in ML, depending on feature characteristics for tree construction [22][23][24].…”

A Survey of Underwater Acoustic Data Classification Methods Using Deep Learning for Shoreline Surveillance

“…When it comes to computer analysis, early work in this field predominantly comprised time–frequency analyses, such as the use of Fourier transforms to temporal data segments [ 13 ]. In contrast, most recent work is based on the application of deep-learning methods to accomplish this task in an automatic way [ 14 , 15 , 16 ]. Due to the rapid development of this field in recent years, a systematic review of deep-learning methods applied to the classification of underwater acoustic data is a timely and relevant task.…”

“…In order to cope with this issue, and also with the complexity in obtaining sonar data from real maritime missions, besides the use of transfer learning, various data-augmentation techniques are described in the literature (summarised in Section 4.1 ). Early techniques and datasets used in this task have been summarised previously in [ 14 , 15 , 16 , 18 ], and the present survey contributes to this group by collating the main ingredients needed to accelerate the development of this area, namely, an up-to-date account of the current methods, existing datasets, and a summary of techniques commonly used for solving the small sample size problem: data augmentation and transfer learning.…”