In the different contexts where the monostatic acoustic seafloor response (ASR) is employed (i.e., measurements with multibeam or singlebeam echosounders, applications in seabed mapping, theoretical models with physical or heuristic parameters), its intrinsic nature can be ambiguous. In this article, an analysis of the discrepancies in semantics is proposed, and then the ASR theoretical and physical definitions are recalled. In theoretical models, and while it is acquired at sea, the ASR is shown to be considered as a deterministic parameter. However, a large variability in its measurements with echosounders is observed, suggesting ASR to be physically a stochastic variable. These two seemingly contradictory observations are linked analytically based on a point scattering model. The theoretical and deterministic ASR is related to its stochastic measurements as the expected value. Based on this connection and under the hypotheses of seafloor homogeneity and large insonified areas, the monostatic ASR measured by echosounders (i.e., the backscattering strength) is derived as twice the parameter of the Rayleigh distribution. Relationships between this result and seabed characteristics are discussed, along with their implications for echosounder operations.

In extreme tidal environment, occurrences of saltating pebbles have been observed. The ambition to instal hydrokinetic turbines in such environment requires knowledge on the presence of pebbles in saltation in the water column because they can damage the structures. An experimental study is realized in a free-surface flume with no slope and different bed roughnesses as in marine environment. With fast camera the trajectories of hundreds of spherical particles are analysed. Our study deals with saltation in the inertial regime (i.e., large Stokes number) over fixed beds with various roughnesses. In inertial regime, the bed roughness has more influence on the collision process and the trajectory of the particles than for non-inertial motion where viscous forces play a key role. Jump height and length increase with bed roughness, with height increasing quicker than length leading to a more vertical trajectory for higher bed roughness. The vertical restitution coefficient is shown to increase with bed roughness leading to higher jumps. The initiation of the motion is shown to depend on the bed roughness as well. Power laws of the excess shear stress are proposed for jump height and length, taking into account the bed roughness. The dataset and analysis proposed in this study is a key ingredient for developing quantitative models for particle transport. An experimental study is realized in a free surface flume with no slope and different bed roughnesses, in hydrodynamically rough turbulent regime as in many natural environments. Jump height and length increase with bed roughness, with height increasing quicker than length leading to a more vertical trajectory for higher bed roughness. The vertical restitution coefficient is shown to increase as well leading to higher jumps. The motion threshold increases linearly with roughness.image

This paper presents the new RTAC acoustic simulator aimed at embedded use on low power GPUs. The purpose of this simulator is to be used for relocalisation tasks based on front scan sonars for Autonomous Underwater Vehicles in already mapped environments. Relocalisation is to be achieved by registration of the real and estimated sonar images. The RTAC simulator is implemented using the CUDA programming model for a deployment on NVIDIA Jetson. Geometrical accuracy of the simulator was tested by comparing simulated sonar image with their real counterpart acquired in natural environnements. To simulate comparable sonar images, the environment map used in the simulator was generated by a photogrammetry software from optical images acquired at the same location at the sonar images. Results show good geometrical consistency between the real and simulated sonar images and the simulator is able to compute one frame per second on a Jetson TX2 module which is enough for relocalisation purposes.

Seabed mapping and characterization are best performed using several frequencies and several angles of incidence. This is often an issue because of the need to employ different sonars, with distinct frequencies but co-located as much as possible to image the same patch of seafloor. This article presents the design, calibration and field testing of a multiple-frequency single-beam echosounder (SBES), mounted on a mechanical pan-and-tilt head. It uses very high transmitting levels to produce non-linear effects and generate harmonics of a 100 kHz fundamental frequency. PZT transducers are used to transmit high acoustic powers and PDVF transducers enable the reception of scattering levels over a very broad frequency band (for the different harmonics). Tank experiments are used to verify effective harmonic generation. The shock distance (at which harmonics are at their maximum level) is measured as 2 m from the transmitter and recommended as the minimum far-field range. Non-linear transmission losses (distinct from linear losses) are calibrated using a full metal sphere 38.1 mm in diameter and of known frequency response, up to ranges commensurate with the depths expected in the field (<= 30 m). The -3 dB beamwidth varies from 5.8 degrees at 100 kHz to 2.8 degrees at 300 kHz. Harmonics are used to resolve phase ambiguities in detecting seabed depths. Backscattering strengths BS are matched to the Generic Seafloor Acoustic Backscatter (GSAB) model to derive the best-fitting parameters. Field validation took place in the Bay of Brest (France) in May 2016, over three different types of seafloor (namely: sandy mud; gravel; gravelly coarse sand with maerl). Additional in situ calibration was used. The echosounder was pointed at angles from 0 degrees (nadir) to 60 degrees by 5 degrees steps. One of the areas surveyed ("Carre Renard"), commonly used for instrument calibration and comparison with other measurements, showed differences <1 dB at 200 kHz. Videos and photographs of the seafloor were used to ground truth interpretations of the BS curves. The results show that these BS curves measured with the echosounder are relevant for seabed classification and characterization. The different shapes and levels of BS when compared to ground truth are coherent with the Jackson model. The main limit of this prototype of echosounder is the signal to noise ratio, in particular for high frequency harmonics (>= 400 kHz). The in situ calibration is unavoidable because of the non-linear parameter variations with water characteristics (temperature, salinity...). Calibrated BS curves from 100 kHz to 300 kHz can be directly compared to other measurements, for example to calibrate other instruments. (C) 2021 Elsevier Ltd. All rights reserved.

The study of underwater acoustic wave propagation provides solutions to localization and underwater navigation problems. In these cases, simulation can be a powerful tool for a better understanding of acoustic propagation. These simulations are based on models that rely on simplifying assumptions allowing the numerical resolution. Simulation is also used to solve more specific problems in underwater environments. For instance, acoustic source localization using receivers in an underwater scene is still a challenging problem and has both civil and military applications. Classical methods are based on the use of acoustic receiver arrays placed in the environment. Assuming a normal modes model for the propagation, collected data are then processed, for example, by singular value decomposition or matched field processing based approach, which provides probabilistic results. The proposed approach to solve this problem is to use set methods. This method allows enclosing all source positions compatible with the recorded hydrophone signal. In addition, possible sets for source position compatible with each receiver can be intersected to increase the certainty of the source location. Besides requiring a good knowledge of the scene, this method requires simulating the acoustic propagation as well as possible to correctly solve this localization problem.

In this paper, we show that a basic fixed point method used to enclose the greatest fixed point in a Kleene algebra will allow us to compute inner and outer approximations of invariant-based sets for continuous-time nonlinear dynamical systems. Our contribution is to provide the definitions and theorems that will allow us to make the link between the theory of invariant sets and the Kleene algebra. This link has never be done before and will allow us to compute rigorously sets that can be defined as a combination of positive invariant sets. Some illustrating examples show the nice properties of the approach.

We present a deep neural network that relies on facial motion and time-domain audio for isolating speech signals from a mixture of speeches and background noises. Recent studies in deep learning-based audio-visual speech separation and speech enhancement have proven that leveraging visual information in addition to audio can yield substantial improvement to the prediction quality and robustness. We propose to use facial motion, inferred from optical flow techniques, as a visual feature input for our model. Combined with state-of-the-art audio-only speech separation approaches, we demonstrate that facial motion significantly improves the speech quality as well as the versatility of the model. Our proposed method offers a signal-to-distortion improvement of up to 4.2 dB on two-speaker mixtures when compared to other audio-visual approaches.

Seabed maps are based on quantities extracted from measurements of the seafloor‘s acoustic response by sonar systems such as single-beam echo-sounders (SBES), multibeam echo-sounders (MBES) or sidescan sonars (SSS). In this paper, a comparison of various strategies to estimate the backscattering strength (BS) from recorded time-series, i.e. seabed echoes extracted from pings, is presented. The work hypotheses are based on processed data from a SBES designed to be tilted mechanically. Ideal survey conditions are taken into account and the seafloor is supposed to be rough so that BS is assumed to be equivalent to the Rayleigh probability density function parameter. Classical methods such as averaging corrected (sonar equation) backscattered single values over a set of pings to estimate BS are compared to other methods exploiting several time-samples being part of pings. Simulated data is considered to estimate BS in different situations (several estimators, natural/squared values, number of samples and pings). The best estimator to reach a 0.1dB uncertainty is proposed, and a formula governing the number of time-samples and pings needed to reach an accurate BS estimation according to the measurement conditions is derived.

Underwater acoustic communication is a key point for performance improvement in an autonomous underwater vehicle (AUV) swarm. The communication process is essential for improving the AUVs localization accuracy for navigation and is a convenient way for sharing information among the AUVs in a network. The objective of this work, which was developed in the COMET and NEMOSENS projects, is to address the communication process required in a mobile underwater wireless network, with a focus on the proposal of an adaptive physical layer methodology. We discuss about the employed channel access method, the frame structure, and we propose the usage of an adaptive guard interval in order to ameliorate the network usage rate. We explain the physical layer aspect of the communication: the data processing at the transmitter and receiver side. In addition to that, we propose the usage of smart communications among AUVs. We design a method for adapting some physical layer parameters. The proposed approach relies only on the knowledge of the transmission geometry, and it optimizes the number of subcarriers and the cyclic-prefix length of the Orthogonal Frequency-Division Multiplexing (OFDM) system. The obtained results show a performance improvement in terms of bit-error rate when compared with the case of random parameters selection. These results corroborate the benefits of our adaptive parameters approach.

This paper presents the development made around theSeaBot, a new low-cost profiling float design for shallow water. We introduce a simplified dynamical model of the float and propose a state feedback depth controller coupled with an Extended Kalman Filter (EKF) to estimate model parameters. We show experimental results of the depth control that validate the model and the controller. We finally propose a loop design method to build low-cost floats by highlighting key design choices along with design rules.

Associated with multi-autonomous underwater vehicle (AUV) systems, the optical sensors with short working range can be reconsidered, which can provide more detailed and more comprehensive descriptions of the seabed than acoustical sensors and can assist in communications among neighbouring AUVs. The typical achievable tasks for these multi-AUV systems equipped with optical sensors mainly include to study the seabed sediments, search small objects at the seabed, collect colourful biological samples, and monitor the underwater infrastructures, etc. The purpose of this research is to develop a new coordination strategy to make a fleet of AUVs build a geometrical pattern suitable for optical sensing at the seabed. This new coordination strategy includes two parts: designing a predefined pattern (a planar pyramid pattern) and proposing an associated formation control method (an improved asynchronous discrete consensus algorithm with a time-variant digraph via the displacement-based control, inspired from the fish schooling problem) to update the trajectories of AUVs in real-time and reach this pattern from a disorder distribution. Before it is tested at sea with six AUVs, the performance of this new coordination method is verified in the simulation environments constructed in Blender and Matlab respectively. The simulation results show the good convergence of this new coordination method.

Multi-robot system has become a research hotspot because of low demand on the sensors’ accuracy, high reliability, and high efficiency. To put all the robots together, formation control is a crucial problem. In this paper, we propose a local position-based method to plan trajectories and build a pyramid pattern for a fleet of NAOs in the obstacle-free environment by refering to the position-based method and giving an O(n log n) collision avoidance strategy inspired from one graph theory, where the local positions are estimated from optical sensors. To get the local positions, an integrated image processing method is developped. Firstly a mask-base is generated to store the features of NAOs, and a cross-correlation method is introduced to recognize the NAO. Subsequently, the distance and angle models are proposed to get the local information from a single image. Then, a visual compass is introduced to obtain the orientation of one NAO. After the local information exchange by the WiFi communication, a neighbor-check method is put foward to distinguish the homogeneous NAOs (all the NAOs look like the same). Further, a common frame is constructed as an artificial global frame, and straight non-intercrossing trajectories are planned according to the O(n log n) collision avoidance strategy. At last, the performance of our proposed local position-based method is verified by the simulations with up to 15 robots and the indoor experiments with 3 NAOs in a real environment. The convergence of the method has been demonstrated in both obstacle-free and static obstacle environments.

Building an accurate and large digital terrain model (DTM) of the seabed is a key issue in various applications, especially for covert rapid environment assessment using autonomous underwater vehicles (AUVs). New AUV generations are capable of acquiring bathymetry with multiple acoustic sensors: singlebeam echosounder, Doppler velocity log, multibeam echosounder (MBES), interferometric sidescan sonar (ISSS), etc. As these sensors acquire the seabed with different geometries, they can be combined to produce a DTM in shorter time. For example, ISSS can reach a wide swath in shallow water but it shows an information gap at nadir, which is usually covered by completing an additional track. Simultaneously using the MBES to acquire the nadir removes the need for this additional track and reduces the energy consumption of the whole survey. This paper focuses on fusion algorithms to extract best information of the two sensors (MBES and ISSS) to feed DTM production software with optimal bathymetric information. This problem may be solved by taking into account the average information of the two sensors. However, the sensors do not always give accurate information and the average information therefore becomes biased. Another way to tackle the problem is to select a priori information given by the better of the two sensors based on a given geometric parameter (e. g., grazing angle). In this case, when the assumed best sensor fails, the information produced by the second sensor cannot be used to compensate for the erroneous information. Our approach consists in using all the available information and fuses it ahead of producing the DTM. Based on the theory of belief functions, this paper presents a framework to fuse the information coming from the two swath bathymetric sensors (MBES and ISSS). The belief theory, applied successfully to other fields, has been extended to handle the bathymetric information. The reliability and the uncertainty of each sonar are introduced in the fusion process to improve the estimation and the accuracy of the final terrain model. First, simulated sonar data, with perfectly known ground truth, are used to quantitatively assess the performance of the fusion process by comparing DTM obtained with and without fusion. Then, the experimental validation is conducted on actual data, acquired simultaneously by the two sonars systems (Klein K5000 ISSS, Reson 8125 MBES) mounted on the DAURADE AUV. Our evaluation of the fusion method shows significant quantitative and qualitative improvement in the production of the final DTM.

In the underwater mine warfare context, change detection is a principle consisting in comparing a newly sensed seabed area, usually by means of a side scan sonar, to another one that has potentially been sensed several months or years ago. In this paper, we propose an approach to simultaneously register (i.e geometrically align) the reference and the repeated data while detecting new and missing objects between both datasets acquisition. This method is first evaluated on data provided by a simulator based on a model of navigation uncertainty as well as on error sources due to the imaging sonar, in order to assess its robustness against different parameters. We also provide results on datasets acquired at sea and demonstrate its efficiency to solve the change detection problem.

In this paper, we present a new method for bracketing (i.e. characterising from inside and from outside) backward reach set of the target region of a continuous-time dynamical system. The principle of the method is to formalise the problem as a constraint network, where the variables are the trajectories (or paths) of the system. The resolution is made possible by using mazes which is a set of paths that contain all solutions of the problem. As a result, we will be able to derive a method able to compute a backward reach set for a huge class of systems without any knowledge of a parametric Lyapunov function and without assuming any linearity for our system. The method will be illustrated in several examples.