Toshihiro Maki

Navigation Method for Underwater Vehicles Based on Mutual Acoustical Positioning With a Single Seafloor Station

In this paper, we propose a novel navigation method for underwater vehicles based on a single seafloor station (SS), with which the vehicles can estimate their positions and orientations with respect to the SS without the need of expensive inertial navigation system or time-consuming calibration. This method is suitable for near-seafloor applications requiring real-time and accurate positioning, such as seafloor imaging and sampling. The method is also suitable for autonomous underwater vehicles (AUVs) since no other external aid is necessary other than SS. The key idea is to utilize mutual acoustical measurements between the vehicle and the SS. Simply explained: 1) the vehicle starts by interrogating the SS acoustically and measures the range between the two units as well as the bearing of the SS with respect to the vehicle in the vehicle reference frame; and then, 2) the SS computes the bearing of the vehicle with respect to the SS and transmits this information back to the vehicle using a similar acoustical device. By combining this information and inputting it into a nonlinear filter structure that includes measurements of the vehicles ground velocity and yaw angular velocity, the vehicle computes its position and heading estimates. A pair of acoustical devices named acoustical localization and communication (ALOC) devices that can communicate and calculate their relative positions have been built. Sea trials were carried out in October 2011 using the AUV Tri-Dog1 (TD) and a trial SS at Kagoshima Bay in Japan. The AUV successfully navigated around the SS based on the measurements of the ALOC device mounted on both the AUV and the SS. The performance of the method was verified through simulations based on the experimental results.

Performance analysis on a navigation method of multiple AUVs for wide area survey

A navigation method of multiple autonomous underwater vehicles (AUVs) is proposed in this paper, and a performance analysis on position estimation for a wide area survey is given. Alternating roles of the AUVs between Moving and Measuring Role (MMR) and Landmark Role (LR) is the key idea of the proposed method. AUVs in the MMR move based on the AUVs in the LR, while AUVs in the LR keep their positions on the seabed to act as landmarks for the moving AUVs. The two groups of AUVs can observe in a wide area, while maintaining a good position estimate, by alternating their roles. AUV position is estimated by mutual acoustic measurements of relative distance among AUVs and direction among them. In order to achieve a robust position estimate against sensor noises and lack of measurements, these position measurements are fused with other on-board sensors such as Doppler velocity logger (DVL), fiber optic gyro (FOG) and depth sensor by particle filter. The performance of the method was verified through navigation simulation based on the performance of sensors obtained during various experiments performed in a pool environment and at sea. In the simulation, the proposed method achieved a positioning error considerably smaller than dead reckoning and conventional navigation methods of a single AUV. Each AUV also achieved state estimation in the proposed method, suppressing deviation from its true position and heading due to alternating roles and evaluating outliers of positioning measurements by particle filter. It is evident that, compared with conventional navigation methods, the proposed method is suitable for a wide area survey without sensor drifts and deviation from true state.

Docking method for hovering type AUVs by acoustic and visual positioning

We propose a docking method for a hovering type AUVs at a seafloor station. Although hovering type AUVs are suitable for detailed surveys of local area, docking methods are currently less studied compared with ones for cruising type AUVs. This paper proposes a docking method for hovering type AUVs based on both the acoustic and visual positioning. The proposed method was implemented in the AUV Tri-TON and a trial station in order to verify its performance through tank experiments. The vehicle succeeded in docking at the station 13 times, which accounts for half of the total number of trials. The possible causes of the failures and improvement methods are also discussed.

Motion estimation and mapping by autonomous underwater vehicles in sea environments

This paper describes a framework for motion estimation and mapping of unstructured underwater environment using multiple onboard sensors. The fundamental goal of this work is to estimate the current pose of the vehicle and build a map of its surroundings. This goal is framed with the context of improving autonomous underwater vehicle navigation for undersea explorations. The paper considers Inertial Navigation System (INS), Sector Scan Sonar and a monocular CCD camera for the purpose of mapping and motion estimation. It discusses a robust feature tracking technique from image sequences and thereby estimating the ego-motion. It also discusses about the trajectory recovery from the sonar and fusion of INS and sonar. Results are shown for the real world data set collected by an autonomous underwater vehicle (AUV) in an unstructured underwater environment.

AUV navigation with a single seafloor station based on mutual orientation measurements

We propose an AUV navigation method based on a single seafloor station. The key idea is acoustic mutual orientation and communication. The AUV firstly sends a signal to the station and then receives the reply with a hydrophone array to estimate direction and distance to the station. On the other hand, the station estimates the direction to the AUV by receiving the signal with a hydrophone array, and sends the information back to the AUV. Then, the AUV can estimate its position and orientation at the station-fixed coordinates without drifts. Furthermore, these measurements are fused with other on-board sensors such as DVL, angular rate gyro and depth sensor by particle filter, a probabilistic approach, in order to realize stable positioning robust against sensor noises and lack of measurements. A pair of acoustic device named ALOC (Acoustic LOcalization and Communication) has been built. A series of sea trials were carried out in September 2010 using a testbed AUV Tri-Dog 1 and a seafloor station mounted with ALOC at Kagoshima Bay in Japan. Although orientation measurements made by the station was not sent in real-time, the performance of the method was verified through simulations based on the actual measurements.

State Estimation and Compression Method for the Navigation of Multiple Autonomous Underwater Vehicles With Limited Communication Traffic

This study proposes a state estimation and compression method for navigating multiple autonomous underwater vehicles (AUVs) toward wide area surveys near seafloors. In the proposed method, a moving AUV navigates by referencing a stationary landmark AUV on the seafloor. By alternating the landmark role, all AUVs can cover a wide area while maintaining low positioning errors. The moving AUV estimates the states (positions and headings) of both moving and landmark AUVs by a stochastic approach called a particle filter. When AUVs exchange their landmark roles, they must share their estimated states. However, state sharing is precluded by the low data rate of acoustical communications in underwater environments. To overcome the problem, this paper proposes a state compression method in which AUVs approximate their states by “particle clustering” based on a clustering method (k-means) and a model evaluation method (Akaike information criterion). The compression method enables AUVs to share their states by communicating small amounts of data. The proposed method was evaluated in simulations of two AUVs navigating over a 300 × 300-m2 seafloor area. Throughout the simulation, the proposed method maintained stable positioning and successful state sharing with small communication data size.

AUV Tri-TON — A hover-capable platform for 3D visualization of complicated surfaces

AUV Tri-TON is a hovering type autonomous underwater vehicle developed by the University of Tokyo, launched in 2011. The vehicle was constructed as a testbed under the governmental project to develop instruments to estimate ore reserves in underwater hydrothermal deposits. The vehicles mission is to obtain dense, large-area 3D image of hydrothermal vent fields, in collaboration with a seafloor station. The information will be also used for environmental assessments, mine planning, and educational activities. Although the vehicle is not equipped with an inertial navigation system (INS), the vehicle can estimate its position in real-time with a precision enough for rough photo-mosaicking, owing to the mutual acoustic positioning with the station. The vehicle has two suites of imaging instruments looking forward and downward directions in order to image whole surface of bumpy hydrothermal vent fields. The vehicle has been tested through a series of experiments at tanks and real fields. In April 2012 the vehicle was deployed to the hydrothermal vent field of Kagoshima Bay in Japan and succeeded in observing seafloor with the area of around 200 square meters.

Autonomous detection and volume determination of tubeworm colonies from underwater robotic surveys

Although the vast amount of information collected by AUVs brings significant benefit to oceanographic research, it is necessary to develop methods to analyze the large volumes of data, in order to avoid accumulation of unused information. Automatic data processing and analysis are key technologies necessary to cope with this problem. We propose a robust, automated method for detection and volume determination of tubeworm colonies using visual and geometric features obtained during underwater robotic surveys, on the condition that the position of the sensors are provided. The tubeworm is a characteristic benthos of hydrothermal vent fields. The proposed method achieves robustness against sensor noise by using both geometric and visual features for identification. First, the tubeworm candidates are obtained as a three-dimensional region between the measured bathymetry of the region and an estimation of the seafloor topology without tubeworms. Next, the tubeworm candidates are verified through frequency analysis of corresponding images. The performance of this method was verified using a data set obtained by the AUV Tri-Dog 1 at Tagiri vent field, Kagoshima bay in Japan.

AUV Tri-TON 2: An intelligent platform for detailed survey of hydrothermal vent fields

The AUV Tri-TON 2 was built in 2013 under the governmental project to develop instruments to estimate ore reserves in underwater hydrothermal deposits, after the success of the prototype AUV Tri-TON [1]. The vehicle has two suites of imaging instruments looking forward and downward directions, in order to obtain dense, large-area 3D image of hydrothermal vent fields. The vehicle can follow rugged surface of hydrothermal vent fields at close range of less than 2.0 m. Although the vehicle is not equipped with an inertial navigation system (INS), the vehicle can estimate its position in real-time with a precision enough for rough photo-mosaicing, owing to the mutual acoustic positioning with a seafloor station. The vehicle has a strong ability of real-time path-planning to obtain a full-coverage 3D image of a rough, unknown seafloor in a single deployment [2]. The performance of the vehicle was verified through a series of sea experiments. At the first experiment, the vehicle succeeded in imaging seafloor with the area of 14 × 10 m. Then, the vehicle was deployed to hydrothermal vent field at Irabu Knoll in Okinawa Trough with the depth of 1,840m.

Imaging vent fields: SLAM based navigation scheme for an AUV toward large-area seafloor imaging

This paper proposes a navigation scheme for an AUV to perform large-area imaging of seafloors, especially those of vent fields. The method is based on SLAM (simultaneous localization and mapping) using a profiling sonar and passive acoustic landmarks. Since real-time positioning accuracy of the method is high enough for rough photomosaicing, this method enables an efficient survey with a small overlap and is applicable to seafloors with no visual key for matching. As well as the positioning method, the navigation scheme contains horizontal and vertical path-planning methods, which adapt the vehicles path to unprescribed landmarks and rough terrains. The method was implemented in the AUV Tri-Dog 1 and a series of sea experiments have been carried out at Tagiri vent field, Kagoshima bay in Japan since 2006. In 2007, the AUV succeeded in imaging around 3,000 m2 of the seafloor and taking micro bathymetry by light sectioning. The relative accuracy of the navigation scheme is estimated to be 0.5 to 1.0 m by comparing photomosaics between dives.