Categories of unmanned underwater vehicles and their basic device components

Remotely operated vehicles (ROVs) and autonomous underwater vehicles (AUVs) are well-known kinds of underwater vehicles. Recently, there are also newer categories of underwater vehicles, untethered ROVs (UROVs) and hybrid ROVs (HROVs). UROVs (Aoki et al., 1992) have the feature that the vehicle is only connected to its support ship via a long thin optical fiber cable. The vehicle of an UROV system has its own power supply, in the form of batteries - much like an AUV. An operator controls the vehicle in real-time and has access to high quality real-time video images using high data rate optical communication tools. UROVs have both the advantages of ROVs and AUVs. An HROV (Bowen et al., 2004), one of which is under development at the Woods Hole Oceanographic Institution, is a single vehicle that can perform two different, but related, missions. It refers to the vehicle's ability to do scientific research while tethered to the ship, and also while swimming freely. Traditionally, a separate vehicle is used to conduct long range surveys, while another vehicle performs the close-up work and sampling. The HROV will simply transform between its two modes of operation to accomplish both of these tasks. In this section, cutting edge basic devices, except for those devices used for controlling vehicles and power sources, are described.

a. Buoyancy Materials and Cables

These are fundamental devices for underwater vehicles. In extreme environments, such as in the deepest depths, a developer should use special devices to match the mission. Full depth buoyancy materials have been commercialized but they have never actually been used in real situations at full ocean depth. The HROV project group at WHOI has chosen SeaSpheres, produced by Deepsea Power & Light, as an alternative to syntactic foams made from micro glass balloons. JAMSTEC has developed a new buoyancy material usable at full ocean depth. The prototype was used in the ABISMO system and it successfully withstood a 10,300 m depth deployment in 2008. The specifications of the prototype are a crush pressure of 56 MPa and a specific gravity of 0.63.

Tether cables for underwater vehicles are also a key device for successful development. Many companies have produced underwater cables, except for cables rated for full depth. Kyo (Kyo 1999) used a Kevlar fiber cable for the full depth vehicle Kaiko, but it was broken during retrieval of the Kaiko vehicle in the face of an approaching typhoon (Watanabe 2004). JAMSTEC thus started the development of a new cable using para-aramid fiber with a tensile strength of 350kg/ mm2 in 2005. This rod type aramid fiber does not concentrate stress. The cable (^20 mm x 160 m) consists of this aramid fiber, two coaxial cables, four single wire cables for power lines, cable sheath, and resin. The cable is covered in polypropylene. Specific gravity of the cable is around 1.3 and rupture strength is about 70 kN.

Fig. 1. A prototype of the full ocean depth buoyancy material (left) and the secondary cable made from para-aramid fiber (right).

Thin fiber optic cable and spoolers are used for UROV and HROV systems. Traditional ^0.9 mm single mode fiber (Murashima 2004) or thinner fiber cable (Young 2006) is practically used for underwater vehicles. b. Lights and Cameras

For the observation of marine organisms, seafloor geology and underwater object recognition, the selection and arrangement of lights and cameras are important. The popularity of high definition television (HDTV) cameras and LED lights are causing an increase in availability of underwater video. In addition to high quality camera imaging, there are holographic cameras, laser scanning systems, acoustic imaging systems and so on. Further information on these imaging systems has been reviewed by Kocak et al. (2008). The underwater vehicle PICASSO, developed by JAMSTEC (Yoshida 2007), is equipped with a broadcast quality HDTV camera. This high resolution, high sensitivity camera enables precise observation of plankton beyond that which was possible with traditional NTSC cameras. The increase in resolution means animals can be identified to species rather than genus or simply family in some cases. JAMSTEC has developed an original wideband optical communication system with five interfaces: one HD-SDI, three NTSCs, four RS-232Cs, two RS-485s, and 8-channel parallel I/ O for the vehicle. This system will be discussed later. They installed SONY's compact high definition camera system, HDC-X300K, and an original camera control board with a CAN interface into an aluminum pressure hull. A special coaxial underwater cable with pressure-tight SMB type RF connectors was made for connecting between pressure hulls. HDC-X300 has the following specifications: effective pixels 1440x1080, sensitivity of 2000 lx @ F10, minimum luminance of 0.003 lx @ F1.4, smear level of -120 dB , and signal to noise ratio of 52 dB. Its image sensor system consists of three 1/2" 1.5M-pixel CCDs. Remote control of the focus, iris, and zoom of this camera via the original control board is possible. The HD-SDI output signal the camera is directly transmitted to an on-board system as an optical modulation signal via the optical communication system. The HD-SDI signal, demodulated and output from the on-board system, is connected to both of an HDCAM recorder and an HDTV display. Any movie subjects are lighted using HID lamps (three custom 30 watt lamps diverted from car use) and/or handmade 20 watts LED array lights. Examples of captured HDTV images obtained by PICASSO are shown in Figure 2.

Fig. 2. An examples of an HDTV images taken by PICASSO-1. In this picture, the sponge and crabs are illuminated by a single HID lamp (left).

High power white LEDs, originally developed by Nichia corporation, have become widely used. Many underwater device makers produce underwater LED lights but they may be expensive. A low cost LED array in an oil-filled pressure balanced case is available to use to 11000 m depth. This consists of LEDs, a copper base plate, resistors, an underwater connector, and a 1/2" clear tube (Yoshida 2007b). c. Stereoscopic HDTV Camera System.

Three-dimensional (3-D) television is one application for a stereoscopic camera system. 3-D television would make an effective operation environment for vehicle operators and viewers. There are lots of commercial software and hardware solutions to make and display 3-D images on a television display and a television screen. Miracube C190x produced by PAVONINE INC. for presentations aimed at small groups employs a 3-D expression method called the Parallax Barrier (Meacham, 1986.). This method doesn't need the observer to wear special glasses but only a single user can enjoy 3D vision and only from certain positions. Use of commercial projector systems for 3-D vision uses shutter glasses or polarizer glasses for users. The use of HDTV cameras for 3-D television gives the audience a more realistic experience. The PICASSO-1 vehicle has the capability to deploy a stereoscopic HDTV camera system. The configuration of the camera system is shown in Figure 3. The major part of the system consists of two pressure-tight HDTV cameras (HDR-SR7 made by SONY) and a controller. Each aluminum pressure hull (^170mm x 390 mm; 9 kilograms in air; depth rating of 4,000 meters; acrylic window) includes an HDTV camera, an interface

Fig. 3. System configuration of the stereoscopic high definition television camera system installed in the PICASSO-1 system.
Fig. 4. PICASSO-1 equipped with the stereoscopic HDTV camera system. Two LED light arrays were additionally made for this system and installed on either side.

adaptor, and a DC-DC converter. HDTV images (MPEG4 AVC/H.264) are locally recorded on the internal 60GB hard disk of the HDR-SR7. Figure 4 shows a snap shot of the PICASSO-1 vehicle equipped with this stereoscopic HDTV camera system.

objecl objecl

Cam era C aim era #2

Fig. 5. Camera placement and coordinate system for stereovision.

Cam era C aim era #2

Fig. 5. Camera placement and coordinate system for stereovision.

The other application for the stereoscopic camera system is as an object scale estimation system. By using HDTV cameras for scale estimation, the resolution of the system become threefold compared with a conventional NTSC-based camera system. For measuring the distance to an object and estimating its size using stereovision, triangulation is generally used. In this method a disparity map is prepared. The disparity map is a depth map where the depth information is derived from offset images of the same scene. Figure 5 shows the coordinate system of the camera system for calculation. The disparity (d) between the left and right image points is defined as the difference between v2 and v\. The depth; D is calculated from equation 1,

Where b, f, and d denote base offset, focal length of camera (distance between lens and film), and disparity, respectively. Object size; S is roughly estimated from equation 2, s=h (Avi+Av>) (2) 2d

In this equation, Av and Av2 are the image size on each film. To measure disparity in the camera system, we compute a given pixel location in either the right or left image coordinate frame with a stereo matching technique. Zitnick and Kanade (Zitnick & Kanade, 1999) have developed a better stereo algorithm. For calculation in real time using high definition images, a very high performance computer would be needed, so this calculation will be done after a dive has finished.

d. Inertial Navigation System (INS)

An INS is one of the most important devices for an AUV because an AUV must obtain an accurate position and information on any attitude changes itself. IXSEA's Phins, which is an INS based on a fiber optic gyroscope having a pure inertial position accuracy of 0.6 NM/hour, is widely used with a Doppler velocity log (DVL) in AUVs. A sufficient level of position accuracy is achieved by the aid of an external sensor, a ground referenced DVL. Larsen reported (Larsen 2002) that the Doppler-inertia based dead-reckoning navigation system, MARPOS, has a proven accuracy of 0.1 per cent of the distance traveled for straight-line trajectories. If an AUV equipped with an INS/DVL hybrid system cruises at a high altitude from a seafloor, a DVL cannot measure its velocity. This leads to increase of positioning error. To reduce this error an AUV usually requires an acoustic navigation system and operators set acoustic transponders in underwater positions before deployment of the AUV. In the case of longer range AUV operations, the time period of AUV navigation using pure inertial positioning data becomes long and this means that many transponders must be deployed - usually an untenable solution. From this point of view an INS should have the highest pure inertial position accuracy possible. Ishibashi et. al. have proposed a unique error reducing technique based on a ring laser gyro (Ishibashi 2008). The position error of an INS results from its drift-bias errors, the sources of which are unidentified random noises. They have proposed a method where the axial rotational motion is applied to the INS. They were able to achieve a high pure inertial position accuracy of 0.09 NM/hour by this method.

e. Ultra Short Base Line (USBL) System

Acoustic navigation systems for underwater vehicles are produced by many companies but USBL systems with full depth capability are very rare. Watanabe et. al. (Watanabe 2006) have developed a small USBL system for full depth use. The system consists of two major parts: a USBL transceiver installed on the station and a transponder fixed on the probe. Table 1 shows the specifications of the USBL system. The accuracy of the position is relatively low because the probe position is directly obtained using the station TV camera in their plan. In this system, the M-sequence signal is used as the modulation signal. An original processing unit has been developed using a DSP (Black Fin produced by Analog devices) and an FPGA (Cyclone produced by Altera). The system was tested in the Marianas Trench in 2008.

0 -1

Post a comment