Transformer-Based Robust Underwater Inertial Navigation in Prolonged Doppler Velocity Log Outages

As the exploration of marine resources continues to deepen, the utilization of Autonomous Underwater Vehicles (AUVs) for conducting marine resource surveys and underwater environmental mapping has become a common practice. In order to successfully accomplish exploration missions, AUVs require high-precision underwater navigation information as support. A Strapdown Inertial Navigation System (SINS) can provide AUVs with accurate attitude and heading information, while a Doppler Velocity Log (DVL) is capable of measuring the velocity vector of the AUVs. Therefore, the integrated SINS/DVL navigation system can furnish the necessary navigational information required by an AUV. In response to the issue of DVL being susceptible to external environmental interference, leading to reduced measurement accuracy, this paper proposes an end-to-end deep-learning approach to enhance the accuracy of DVL velocity vector measurements. The utilization of the raw measurement data from an Inertial Measurement Unit (IMU), which includes gyroscopes and accelerometers, to assist the DVL in velocity vector estimation and to refine it towards the Global Positioning System (GPS) velocity vector, compensates for the external environmental interference affecting the DVL, therefore enhancing the navigation accuracy. To evaluate the proposed method, we conducted lake experiments using SINS and DVL equipment, from which the collected data were organized into a dataset for training and assessing the model. The research results show that the DVL vector predicted by our model can achieve a maximum improvement of 69.26% in terms of root mean square error and a maximum improvement of 78.62% in terms of relative trajectory error.

The objective of this thesis is to contribute to Autonomous Underwater Vehicle (AUV) navigation by improving vehicle localisation accuracy when Doppler Velocity Log (DVL) bottom-tracking is unavailable. The Inertial Navigation System (INS) based localisation solution is prone to extreme uncertainties due to double integration of inherent errors within the INS acceleration measurements, unless the solution being externally aided (e.g. velocity measurements using the DVL bottom-track). As a solution for this, an improved model-aided INS localisation technique is introduced, which is complimented with the development of a novel model calibration and new water column velocity estimation method. The techniques established in this project are tested and validated using experimental data from a set of field manoeuvres using a <i>Gavia</i> class AUV and the performance is compared against other commonly used localisation methods. A baseline mathematical model was developed in this work using system identification to predict the motion response of the AUV based on its control commands. However, such models are generally calibrated for low water column velocities and a standard vehicle configuration, and are limited in application for variations in environmental conditions. To address this limitation, a novel model calibration technique was established to field calibrate the parameters within the baseline model to the current operating condition and vehicle configuration. Model calibration improved the results of the baseline model up to 73% when operating in low energy environments and the AUV position can be computed within an uncertainty range of 1.5% of the distance travelled. In comparison, uncertainties of conventional non-bottom-tracking localisation techniques could be up to 10% in similar environmental conditions. A secondary approach is also presented to determine the hydrodynamic coefficients of the mathematical model using Computational Fluid Dynamics (CFD) simulations and captive model experiments when the AUV is operating in complex flow conditions. A non-acoustic method was introduced to estimate the velocity of the water column using the motion response of the AUV. This method is capable of accurately estimating water column velocities in proximity to the AUV (i.e., the water column velocities at the same depth as the vehicle is), which is not typically resolved with existing methods such as acoustic Doppler current profilers. When the mathematical model-aided localisation solution is complimented with water column velocity prediction method, the localisation error is limited to less than 6% of the distance travelled even in extremely high currents (i.e. &gt;2 m s-1 in tested conditions). In such environmental conditions, the uncertainties of other commonly used non-bottom-tracking localisation methods, as tested against in this work, such as DVL water-track mode and unaided INS could be above 30%. One of the key advantages of the proposed localisation technique is that it could be applied to any torpedo shaped AUV (for example, platforms such as <i>REMUS, Iver, Bluefin, Explorer</i>, etc.) of any configuration by simply conducting a set of established field manoeuvres to identify its mathematical model parameters. Further, additional sensors beyond a typical AUV navigational payload (i.e. global positioning, accelerometers and gyroscopes) are not required to implement this technique in an AUV. The localisation technique developed in this thesis is capable of improving the motion control and navigation solution of the AUV in the absence of DVL bottom-tracking, which is critical for the expansion of vehicle performance in extreme environmental conditions.

We, Japan Agency for Marine-Earth Science and Technology (JAMSTEC), are developing an autonomous underwater vehicle (AUV) whose main mission is monitoring a site at the sea bottom for the carbon dioxide capture and storage (CCS). The AUV cruises very near the sea bottom, and is equipped with chemical sensors in order to detect escape of CO2 from sub-bottom. Of course, the position information of the AUV is critical information for the monitoring. In this paper, a conceptual design of navigation of the AUV is described. Recently, navigation of AUV is implemented by integrating multiple navigation devices including inertial navigation system (INS), Doppler velocity log (DVL), depth sensor, acoustic navigation system, and others. The AUV under construction will be equipped with these navigation sensors, and will integrate those sensors’ outputs to navigate herself. In order to measure the absolute position of the AUV the acoustic method is one of fundamental technique. At the first step of development of the AUV, three acoustic methods are considered to adopt. The three methods are super short baseline (SSBL) method which is a tracking from support ship or other surface station, long baseline (LBL) which is navigation based on preplaced acoustic transponders, and virtual LBL (VLBL) which is navigation based on only single transponder. These acoustic methods are integrated with the navigation result of INS, depth sensor, and DVL. The three methods are used in each appropriate case. Which feature of observation is desired simplicity, accuracy, or independence from support ship and time efficiency? The acoustic method is influenced by environment, and also output of other sensors is depending on the environment, for example the DVL miss the data when the terrain is with many up-hills and down-hills. The integration or filtering parameters of the navigation should be adjusted depending on the influential environmental factor.

Recent development and commercialization of relatively portable and inexpensive autonomous underwater vehicles (AUVs) have led to new methods for making hydrodynamic, chemical, and biological measurements in near-coastal, lake, and estuarine systems. Equipped with Doppler velocity log (DVL) technology to allow for deadreckoning navigation, this new class of AUVs has the potential to log not only vehicle motions, but water velocities as well. In this article, we assess the performance of the DVL acoustic Doppler current profiler (ADCP) equipped on the REMUS AUV (Hydroid) in three different environments: a tidally forced lake; a wave-forced, near-shore fringing coral reef; and a gently sloping continental shelf. In all three data sets, the water velocities measured by the AUV DVL compare favorably with measurements from nearby stationary acoustic Doppler current profilers. Nevertheless, the water velocity component measured parallel to the vehicle's tracks exhibited a bias in the direction of AUV motion, consistent with a previously reported bias found for the ship-mounted ADCP application in low-scattering environments. The fringing reef data suggest that DVLs are capable of partially resolving wave motions.

The Monterey Bay Aquarium Institute (MBARI) has conducted high-resolution multibeam bathymetry, sidescan, and subbottom sonar surveys in the Gulf of California using an autonomous underwater vehicle (AUV) as part of a three-month, two-ship oceanographic expedition. The seafloor mapping AUV D. Allan B. was operated from R/V Zephyr, achieving 22 successful 17-hour missions during March and April 2012. The 1-m lateral resolution seafloor maps and subbottom profiles were used to guide subsequent ROV Doc Ricketts dives made from MBARI's R/V Western Flyer, which combined visual surveys, sampling, and in situ experimentation by several different science groups. The vehicle is a modularized Dorado Class 54 cm (21 inch) diameter, 5.18 meter (17 feet) length, 733 kg (1616 pounds) air weight. The interior of the AUV floods with water when launched; electronics are contained in either one-atmosphere hermetic housings or in pressure-compensated enclosures. The AUV is ballasted to be about 3.63 kg (8 pounds) positive in seawater. Flotation is accomplished with 6000 meter rated syntactic foam. The D. Allan B. has three mapping sonar systems onboard: a Reson 200 kHz multibeam echosounder, an Edgetech FS-AU Sonar Package which comprises a 100 kHz chirp sidescan, a 410 kHz chirp sidescan, and a 2-16 kHz chirp subbottom profiler. With the onboard Lithium Polymer battery packs, the Mapping AUV has an endurance of approximately 17 hours at a speed of 1.5 meters per second, providing a range of 55 to 85 kilometers. The mapping AUV is equipped with a Kearfott inertial navigation system (INS) with Doppler velocity log, a Paroscientific pressure sensor, and an ultrashort baseline (USBL) and acoustic modem for communications. It also has an Iridium Satellite Short Burst Data modem, a drop weight, radio direction finder, Homer Pro and visible strobe location aids. Propulsion is accomplished with a duct-protected articulated propeller which provides 52 Newtons (12lbf) of thrust at 300 rpm as well as elevator and rudder functions. Turning diameter is less than 20 meters; climb/dive rate is approximately 30 meters per minute. Launches and recoveries were made from the RV Zephyr, a 24m (79 foot) former pilot boat that operates with four crew and four science personnel. The Zephyr is outfitted with a Brook Ocean Technologies J-Frame crane for AUV launch and recovery. Once the AUV is in the water, the mission is sent from a computer on the Zephyr to the AUV over a wireless Freewave Ethernet link. At the start of the mission, the AUV spirals down to the start-of-line while being position-updated with GPS data from the Zephyr aided by the USBL system. A Bottom Pressure Recorder used to provide in situ tide data for bathymetry correction was deployed prior to all AUV missions, and recovered afterwards, using a Benthos Recoverable Transponder mounted to an MBARI-designed frame. The seafloor mapping data were processed using MB-System, an open source software package developed at MBARI and the Lamont-Doherty Earth Observatory. The first half of the mapping AUV surveys focused largely on four transform faults in the central and northern gulf. The combination of high-resolution bathymetry and subbottom profiles enabled the identification of active fault traces and zones of subsurface gas and hydrate. The second half of the Zephyr expedition focused on the Alarcon Rise, a volcanic spreading center. The new surveys provide the first 1-m scale bathymetry coverage of the neovolcanic zone along an entire ocean spreading center segment. The mapping data also identified the locations of several fields of hydrothermal vent chimneys, two of which were confirmed to be currently active by ROV dives.

A Doppler velocity log (DVL) provides a three-dimensional representation of the velocity of an underwater vehicle to the vehicle’s DVL-aided inertial navigation system (INS). It is a key element in determining the precision of the INS. This paper presents an algorithm for reducing the error in the DVL as a way of improving the navigation of an autonomous underwater vehicle (AUV). We also investigate and analyze the effect that the seafloor’s deviations have on DVL measurement error. Finally, the performance of the proposed algorithm is examined through both AUV simulation and experimentally. The results show that the proposed technique can indeed effectively correct DVL errors.

Underwater navigation methodology based on intelligent velocity model for standard AUV

This paper reports a preliminary study for use of a fully dynamic vehicle process model in combined underwater communication and navigation (cooperative navigation) of underwater vehicles equipped with an acoustic modem, attitude, and depth sensors, but lacking a Doppler velocity log (DVL), and a surface vehicle equipped with an acoustic modem and GPS. We report both simulation and at-sea experimental trials with the JHU Iver3 autonomous underwater vehicle (AUV). The case of underwater vehicle navigation without a DVL is of interest in several use-cases including (a) small and low-cost underwater vehicles for which DVLs may be impractical or infeasible due to their size and cost and (b) for missions in which the vehicle's altitude above the sea floor (or depth beneath overhead ice) exceeds the DVL acoustic bottom-lock range. To the best of our knowledge, all previous studies on cooperative navigation have reported use of a kinematic process model, which works well in the presence of frequent, high-accuracy velocity measurements, as is the case when the vehicle is equipped with a DVL. This preliminary study suggests that the dynamical process model may offer a significant advantage over the purely kinematic model in the absence of frequent, high-accuracy velocity measurements, as is the case when the submerged vehicle is not equipped with a DVL.

This paper proposes a new method for cruising-type autonomous underwater vehicles (AUVs) to track rough seafloors at low altitudes while also maintaining a high surge velocity. Low altitudes are required for visual observation of the seafloor. The operation of AUVs at low altitudes and high surge velocities permits rapid seafloor imaging over a wide area. This method works without high-grade sensors, such as inertial navigation systems (INS), Doppler velocity logs (DVL), or multi-beam sonars, and it can be implemented in lightweight AUVs. The seafloor position is estimated based on a reflection intensity map defined on a vertical plane, using measurements from scanning sonar and basic sensors of depth, attitude, and surge velocity. Then, based on the potential method, a reference pitch angle is generated that allows the AUV to follow the seafloor at a constant altitude. This method was implemented in the AUV HATTORI, and a series of sea experiments were carried out to evaluate its performance. HATTORI (Highly Agile Terrain Tracker for Ocean Research and Investigation) is a lightweight and low-cost testbed designed for rapid and efficient imaging of rugged seafloors, such as those containing coral reefs. The vehicle succeeded in following a rocky terrain at an altitude of approximately 2 m with a surge velocity of approximately 0.8 m/s. This paper also presents the results of sea trials conducted at Ishigaki Island in 2017, where the vehicle succeeded in surveying the irregular, coral-covered seafloor.

In order to improve the positioning accuracy of two leader autonomous underwater vehicles (AUVs) in underwater cooperative navigation formation, a decentralized cooperative localization algorithm based on adaptive cubature Kalman filter (ACKF-DCL) for strapdown inertial navigation system (SINS)/Doppler velocity log (DVL) integrated navigation system is proposed. Firstly, the kinematic equations of the SINS are used to establish the direct filtering approach model for integrated navigation. Secondly, the Sage-Husa adaptive filtering algorithm and the sequential Kalman filtering (SKF) method are introduced into the nonlinear cubature Kalman filter (CKF) to adjust adaptively the measurement noise and maintain the positivity of the covariance matrix. Finally, the cross-correlation covariance between two AUVs is derived according to the characteristics of a decentralized data structure, so as to track the inter-AUV dependencies. And the acoustic ranging measurement is introduced to achieve relative measurement update. The lake trial demonstrates the proposed algorithm&#x2019;s effectiveness and superiority, and the results show that the algorithm can effectively track the measurement noise covariance matrix and improve the positioning accuracy of both AUVs by position correction.

To improve the navigation accuracy of AUV (autonomous underwater vehicle), SINS (strapdown inertial navigation system), GPS (global positioning system), DVL (Doppler velocity log) and TAN (terrain aided navigation) are adopted in the AUV integrated navigation system. The mathematic model of the AUV integrated navigation system and the observation model of the chosen navigation sensors are built according to the system simulation experiments data. An improved filter based on RBF neural network for adjusting the information sharing factors is designed and implemented in the AUV integrated navigation system. Simulation experiments are carried out according to the mathematic model. It can be concluded from the simulation experiments that the navigation accuracy is improved substantially with the multiple sensors and federated filter in case that colored noise is engaged. The novel integrated navigation system is effective in prohibiting the divergence of the filter and improving fault tolerance ability and it greatly raises the precision of the navigation accuracy for the AUV integrated navigation system.

International audience

In recent years, precious observations by autonomous underwater vehicle (AUV) have been performed. Tracking and navigation of AUV are important for precious observations. In Japan agency for Marine-earth Science and Technology (JAMSTEC) the observations with AUV ldquoURASHIMArdquo have been implemented in deep ocean area. AUV URASHIMA is navigated by an inertial navigation system (INS) mainly. The absolute position of AUV is tracked on the support vessel with super short base-line (SSBL) method of acoustic positioning. And sometime the own position in AUV is updated by the SSBL estimation result to correct the drift error of INS. However the SSBL is not so accurate. So the position correction includes the SSBL error and the position correction is performed drastically. In this paper, a method for accurate tracking of AUV and rapid correction of the INS status is proposed. The tracking is mainly achieved by SSBL on the support vessel. Additionally the INS data is transmitted from AUV including the acoustic positioning signal. The INS output of relative position to past own position is very accurate within short time. The INS data is used for tracking. Furthermore the position result estimated on the vessel is transmitted to AUV with next interrogation signal. And the own position in AUV is rapidly updated with accurate estimation. Some digital data including INS data is modulated with phase shift keying (PSK) method and direct sequence spread spectrum (DSSS) technique for transmission with acoustic positioning signal. The proposed method was considered by simulation. And good results were obtained.

A novel intelligent adaptive Kalman Filter for estimating the Submarine's velocity: With experimental evaluation

This paper describes an approach that combines the navigation data given by a Doppler Velocity Logs (DVL), the MTi Motion Reference Unit (MRU) and a Mechanically Scanned Imaging Sonar (MSIS) as a principal sensor to efficiently solve underwater Simultaneous Localization and Mapping (SLAM) problem in structured environments such as marine platforms, harbors, or dams, etc. The MSIS has been chosen of its capacity to produce a rich representation of the environment. In recent years, to solve the SLAM Autonomous Underwater Vehicle (AUV) problem, very few solutions have been proposed. Our contribution has introduced a method based on the Nonlinear H-infinity filter \((NH\infty )\) to solve the SLAM-AUV problem. In this work, the \(NH\infty \)-SLAM algorithm is implemented to construct a map in partially structured environments and localize the AUV within this map. The data-set used in this paper are taken from the experiments carried out in a marina located in the Costa Brava (Spain) with the Ictineu AUV which is necessary to test different SLAM algorithms. The validation of the proposed algorithm through simulation in offline is presented and compared to the EKF-SLAM algorithm. The \(NH\infty \)-SLAM algorithm provides an accurate estimate than EKF-SLAM and good results were obtained.