Autonomous Underwater Environment Perceiving and Modeling: An Experimental Campaign With FeelHippo AUV for Forward Looking Sonar-Based Automatic Target Recognition and Data Association

Maximum A Posteriori estimation for AUV localization with USBL measurements

In recent years, convolutional neural network (CNN) has been increasingly considered as a promising technology for military and homeland security applications. The fusion of CNN and Support vector machine (SVM), a popular traditional machine learning approach, has received intensive attention in the field of synthetic aperture radar (SAR) automatic target recognition (ATR). This paper, firstly, discusses the effects of some preprocessing and image enhancement methods on the performance of SAR ATR, starting with the pre-trained AlexNet in a transfer-learning based approach. Secondly, the architecture of AlexNet is modified to form a new model suitable for SAR ATR. Finally, we propose a hybrid model associated with the success of the learning feature of our CNN model and the ability of SVM to process high-dimensional dataset effectively. To evaluate the proposed method, experiments are performed on the Moving and Stationary Target Acquisition and Recognition (MSTAR) public database. The comparative results demonstrate that these preprocessing and enhancement methods prior to the deep-learning process are not necessary since the feature representation ability of AlexNet is already powerful. Furthermore, experimental results on the benchmark MSTAR dataset illustrate the effectiveness of the proposed new model. On classification of ten-class targets, the commonly used translation augmentation for training data has been performed. By combining the CNN and SVM, the classification accuracy percentages can be slightly improved for our proposed new model.

The sensory processing system and its relationship to the world model are described. Both the sensory-processing and world-modeling systems are hypothesized to be hierarchical, with a tight coupling between sensory-processing and world-modeling modules at each level of the hierarchy. At every level, world model predictions are compared with sensory observations. Correlations and differences are used to improve estimated values stored in the world model. The world model consists of both symbolic and iconic representations and of mechanisms for transforming from one to the other. The sensory-processing system consists of iconic images, coordinate transformers, and mechanisms for comparison, correlation, recognition, and detection. The close coupling between sensory processing and world modeling produces the phenomenal capacity of an intelligent system to recognize objects, detect events, and perceive what is important for successful behavior in the natural world. >

In light of the prevailing approach in which data from side-scan sonar (SSS) from Autonomous Underwater Vehicles (AUVs) are primarily processed and visualized post mission, failing to meet the requirements in terms of timeliness for on-the-fly image acquisition, this paper introduces a novel method for real-time processing and superior imaging of navigation strip data from SSS aboard AUVs. Initially, a comprehensive description of the real-time processing sequence is provided, encompassing the integration of multi-source navigation data using Kalman filtering, and high-pass filtering of attitude and heading data to exclude anomalies, as well as the use of bidirectional filtering techniques within and between pings, ensuring real-time quality control of raw data. In addition, this study adopts the semantic segmentation Unet network for automatic real-time tracking of seafloor lines, devises a real-time correction strategy for radial distortion based on historical echo data, and utilizes the alternating direction multiplier method for real-time noise reduction in strip images. With the combined application of these four pivotal techniques, we adeptly address the primary challenges in real-time navigation data processing. In conclusion, marine tests conducted in Bohai Bay substantiate the efficacy of the methodologies delineated in this research, offering a fresh paradigm for real-time processing and superior visualization of SSS navigation strip data on AUVs.

Detection, classification and identification of underwater objects are both complex and essential tasks for perception and navigation of Autonomous Underwater Vehicles (AUVs) and Remotely Operated Vehicles (ROVs). AUVs need to detect obstacles to avoid them and need to identify certain objects in tasks such as Automatic Target Recognition (ATR) or in manipulation tasks in the case of ROVs. Leveraging a dataset published by the Robotics for Asset Maintenance and Inspection (RAMI) competition for the 2023 Virtual competition, in this work, we detail our machine learning-based solution for the visual detection, classification and identification of several objects including buoys of different colours, markers with numbers, pipes and ring poles among others. The results show excellent performance in training dataset and very good performance for detection and classification in the unknown test dataset showing the goodness of the methods in harsh environments with low visibility. Identification remains a challenging task especially for not previously seen data and requires further work including synthethic data generation.

A synthetic aperture radar (SAR) automatic target recognition can effectively improve the utilization efficiency of SAR image data. In order to improve the generalization and accuracy of SAR image target recognition, a convolutional neural network model is proposed to be applied to SAR image target recognition. Firstly, the network structure is designed for the characteristics of SAR image. Secondly, the introduction of dense block network structure improves the generalization performance of the model. Finally, Dropout reduced the computational complexity of the network and improved the generalization performance of the model. Experimental data were obtained from the United States Moving and Stationary Target Acquisition and recognition database. Experimental results of 10 types of target recognition showed that the overall recognition rate of the improved convolutional neural network is 99.18%. The convolutional neural network model proposed in this study improves the accuracy and generalization of the network, and is an effective method for target recognition of SAR images.

Hydrophone, an acoustic sensor, used to pick up acoustic signals generated by different objects present in underwater. MEMS Vector Hydrophone is a recent advancement in the field of underwater acoustic sensors that provides the direction of the incoming acoustic signals. As the main sensing elements of these hydrophones are in micrometers hence light weight, which makes them useful for deployment in Autonomous Underwater vehicles (AUVs) and Remotely Operated Underwater Vehicles (ROVs). This paper presents simulation, modeling and design optimization of four beam MEMS Vector Hydrophone, using design modeler of ANSYS Workbench. This paper presents design optimization, stress, frequency, harmonic and electro-mechanical analysis of MEMS-based four beam Vector Hydrophone. Using ANSYS software design is optimized by changing important parameters to improve design and sensitivity that can be chosen for fabrication. Static structural analysis is performed on the structure for getting an optimized location for placing piezoresistors over cantilever beams. For better understanding of structural vibrations that can be induced in the Vector Hydrophone structure, the resonant frequencies of the structure were identified and quantified. The Receiving Voltage Sensitivity is −196.73 dB (0 dB = 1 V/µPa) which is further improved to −184.86 dB by design optimization (Xue et al. in Microelectron J 38:1021–1026, 2007 [1]). By incorporating pre-amplifier having a gain of 100, Voltage sensitivity further can be enhanced up to −156.73 dB (Hang et al. in Nano-Micro Lett 6:136–142, 2014 [2]).

One of the key aspects in designing an Autonomous Underwater Vehicle (AUV) simulation framework is sensor modeling. This paper presents specifically the underwater sonar sensor modeling structure used in the proposed AUV simulation framework. This sensor model covers the mathematical aspects from the field of acoustics which mimics real world sensors. Simplified sonar signal models are widely used however rarely discussed in the literature. Based on this designed simulation framework, simple scenario using different sonar configuration is shown and discussed. This paper shows the formulation of a typical side-scan sonar with emphasis on the assumptions which leads to the simplification of the sonar model. The sonar sensor model is built based on a developed AUV test-bed which was done previously in the University of Adelaide.

Modeling of the Ballard-Mark-V proton exchange membrane fuel cell with power converters for applications in autonomous underwater vehicles

We describe how an autonomous underwater vehicle (AUV) engaged in the task of mosaicking an unknown region of the ocean floor recognizes "essential" occlusions in its sensor image. The sensor on board the AUV is either a downward pointing camera or a side-scan sonar. Essential occlusions, which are created by irregularities in bathymetry, force the AUV to make diversions from its regular mosaicking path in order to prevent gaps in the mosaic. We classify features on the ocean floor which give rise to visual occlusions into three categories: knolls, furrows and dimples. Geometric conditions are derived that enable the AUV's camera to distinguish between these occlusions. We show that only furrows and dimples give rise to essential occlusions. We discuss extensions of the geometric conditions to shadows in side-scan records. There, however, the conditions are not as straightforward, and their implementation constitutes work in progress. An example is provided, which exhibits the modifications required in a visual mosaicking procedure in a complex environment.

The performance of an automatic target recognition (ATR) system in the context of naval mine detection is severely affected by the underwater environment. Especially in regions with the presence of sand ripples or mine-sized stones the number of false alarms can become unacceptable high, if the detection algorithm does not account for the type of seafloor. Therefore, a robust way of discriminating between the three most important types of seafloor in the context of mine countermeasures (MCM), flat bottom, rocky bottom and sand ripples, is needed. In this paper five handcrafted features for seafloor classification from the literature are analysed and compared to the performance of a convolutional neural network (CNN) that learns the relevant features from the training data. The evaluation data was collected with an autonomous underwater vehicle (AUV) equipped with an synthetic aperture sonar. Experiments showed that the CNN approach outperformed a support vector machines classifier with the handcrafted features, albeit slightly, with a classification rate of 98.7% over 9420 examples.

The author describes autonomous underwater vehicle (AUV) developments and applications world-wide, AUV development in China, and AUV trends. Besides developments in China, those in the USA, Canada, the United Kingdom, France, Germany, Japan, and Russia are discussed, such as deep ocean modelling, ocean bottom exploration, ocean bottom sampling, deep ocean work platform, underwater survey, and so on. A testbed facility is described. AUV research topics are outlined. >

Recent advances in transducer and computing technology have pushed the concept of continuous active sonar (CAS) or high duty cycle sonar as an area of interest for application to antisubmarine warfare. Unlike conventional pulsed active sonars, CAS processing aims at detecting echoes while transmitting with a nearly 100% duty cycle. This paper describes the signal processing chain developed at the Centre for Maritime Research and Experimentation (CMRE), La Spezia, Italy, for real-time CAS processing within an autonomous networked multistatic sonar system. The algorithm uses subband processing, which can potentially provide a higher target update rate than the traditional pulsed active sonar, while maintaining the same search radius. The higher rate of contact information can improve target tracking performance. Performance results are given from the COLLAB-NGAS14 sea trial where the CAS processor was deployed on two autonomous underwater vehicles (AUVs) acting as receivers of the CMRE experimental multistatic demonstrator. Results show the feasibility of the CAS concept in littoral scenarios, using AUVs as real-time receivers.

For decades, broadband sonar has been a key tool in the underwater detection and classification of man-made targets. The DoD’s recent focus to clean up discarded WWII-era munitions has pushed a flurry of research and advancements in the areas of target scattering, environmental noise modeling, signal processing, and automatic target recognition (ATR). It is the interdisciplinary nature of this problem that makes it incredibly challenging, namely, the fact that the ability to classify an object is heavily influenced by the operational choices (grazing angle and vehicle altitude), surrounding environmental factors (target type, burial state, sea state, and bottom topography), and the processing tools (data products, feature-set, and ATR training/testing). This talk will outline the key steps underlying the process of trying to detect and classify underwater unexploded ordnance (UXO). It begins by pushing raw time domain data, which can be generated either experimentally or via models, through a set of signal processing tools to create data products (acoustic color and SAS images). These data products are fed through feature extraction processes, followed by ATR algorithms, in order to arrive at the final classification of an object. Examples to be presented include objects of interest to the remediation of underwater UXO. [Research supported by SERDP and ONR.].

Localization and tracking are basic services for the applications of autonomous underwater vehicles (AUVs). Although they are closely related, the localization or tracking is usually studied independently. This paper addresses a joint localization and tracking issue for an AUV, subject to asynchronous clocks in physical channel and state disturbances in control channel. We first develop an asynchronous localization algorithm to estimate the position of an AUV, wherein the clocks on AUV and buoys are not required to be synchronized. With the estimated position, a model-free tracking controller is designed for an AUV to track the reference trajectory. Particularly, the time-delay estimation and dynamic gain strategies are adopted to reduce the influence of state disturbances. Moreover, the convergence conditions and Cramer–Rao lower bounds for the localization algorithm are derived. For the tracking controller, Lyapunov functions are constructed to analyze the stability, through which the sufficient conditions for asymptotic stability are presented. Finally, simulation and experiment results are given to show the effectiveness of the proposed method. It is demonstrated that the asynchronous localization algorithm in this paper can effectively eliminate the impact of the clock skew and offset. Meanwhile, the tracking accuracy can be significantly improved.