Feature ArticleAutonomous AUV Navigation In a Partially Known Environment
In April 2009, the Korea Institute of Robot & Convergence (KIRO), formerly Pohang Institute of Intelligent Robotics, (PIRO), launched a project to develop a test-bed AUV called P-SURO (PIRO-Smart Underwater Robot) to demonstrate various AUV technologies, such as vision-based underwater localization, obstacle avoidance and underwater simultaneous localization and mapping (SLAM).
Four SeaBotix Inc. (San Diego, California) SBT150 thrusters, two vertical and two horizontal, are mounted on the vehicle. Throughout its underwater mission, the vehicle is always forced to keep zero pitch angles. With this kind of thrust configuration, the vehicle’s horizontal and heavy motions can be easily decoupled.
There are three communication channels connecting the vehicle and the surface unit. Ethernet cable is used in the early stage of development and to upload and download files. On the surface, radio frequency is used to exchange information, while LinkQuest Inc.’s (San Diego, California) UWM2000H acoustic modem is used underwater. Tritech International Ltd. (Aberdeen, Scotland) Super SeaSpy camera is mounted in front of the vehicle and used for vision-based localization. A total of three Tritech Micron Echosounders are used for environment recognition, including obstacle detection.
Since August 2010, various water-tank tests have been carried out to demonstrate the developed AUV technologies.
The P-SURO AUV and its open frame.
P-SURO’s experimental tests are carried out mostly in a 12-by-8-by-7-meter concrete cuboid water tank at KIRO. The underwater acoustic link is vulnerable in the tank, so the vehicle must possess a relatively high level of autonomy.
P-SURO was designed to be compact at 1.1 meters length, 0.5 meters width and 0.3 meters height. Two pressure hulls, each for a battery system and electronics system, are designed for maintenance and reliability. To increase its hydrodynamic mobility, the open frame of the vehicle is wrapped by two pieces of fiber-reinforced plastic shell.
There are only two horizontal thrusters to steer the vehicle at 3 degrees of freedom (DOF) horizontal motion. From a control point of view, this is a typical underactuated system. How to design a stable tracking scheme for this system has been one of most interesting research topics in the nonlinear control community of the past decades.
An InnaLabs (Dublin, Ireland) AHRS (attitude and heading reference system) is mounted on the vehicle. This is a microelectromechanical systems device, with three axes magnetometers installed to measure the geomagnetic field vector to correct the drift of calculated heading. However, KIRO’s water tank is surrounded by a steel structure, and there is a considerable magnetic-field distortion in the tank environment. Consequently, AHRS could not output proper attitude value.
However, there is a one-axis InnaLabs CVG-25 gyroscope, with random walk less than 0.05°/√h, mounted on the vehicle’s horizontal plane. By integrating this gyro output with an online calibration process, a heading value can be obtained with less than 1 degree of drift.
Windows Embedded CE 6.0 is applied in near real time to operate the embedded system, which further consists of three single board computer-based modules for vision, navigation and control. The modules connect to each other through TCP/IP ports. The real-time layer of the software frame is a thread-based multitasking structure.
According to their accessing mechanism, sensor devices can be classified into two types of active sensors that automatically output measurement and a passive one that only outputs after being requested. The passive sensor measurements as well as analog sensors use a timer routine, while individual interface threads are designed for each of the active sensors.
Visual localization methods usually can be classified into two types. One type is based on the natural feature points of the environment, and the other type uses artificial landmarks. The KIRO water tank is surrounded by flat concrete walls, and it is difficult to extract specific feature points, so artificial landmarks were applied for vision-based localization. To continue this article please click here.
Ji-Hong Li received a Ph.D. in electronics engineering in 2003. He has worked previously for the Maritime and Ocean Engineering Research Institute and joined the project developing the 6,000-meter depth-rated Hemire ROV. His research interests are control and guidance of underwater vehicles. He moved to KIRO in 2009, where he is the director and principal researcher at the Underwater Robotics Lab.