Autonomous Underwater Vehicles of The University of Tokyo

Figs and Tables are not yet.
Proc. of IARP Workshop at Monterey, 1994
Tamaki Ura
Institute of Industrial Science, University of Tokyo
4-6-1, Komaba, Meguro, Tokyo 153-8505 Japan
Tel:+81-3-5452-6487 Fax:+81-3-5452-6488

ABSTRACT

Institute of Industrial Science of the University of Tokyo started research and development of autonomous underwater vehicles at 1984 and has constructed three vehicles which can be operated at actual sea. The PTEROA150 is a prototype of cruising type AUVs. The ALBAC is a shuttle type AUV which measures water column to and from the bed. The TWIN-BURGER is a versatile test bed for advanced research on AUVs. On the basis of these technology, a new project called "R1 Project" was started from 1990 to construct a long endurance vehicle which can be operated for one day. R1 Robot would be completed in 1995.This paper introduces the state of art of these vehicles and results of trials.

1. INTRODUCTION

In order to understand what is happening under the surface of the ocean, we should operate underwater vehicles in the environment where vast expanse and hostile depth stands in their way. Though manned submersibles and ROVs are powerful tools, it is clear that they are not all-powerful. It can be said, at present, that there is no technology which overcomes the above difficulties. By analogy with airplanes which include Jumbo-jets, fighters, and helicopters, various types of underwater vehicle should be created which conduct specified performances under the sea in their own way.
In order to be competitive to manned submersibles and ROVs, Autonomous Underwater Vehicles (AUVs), which are anticipated to be powerful tools for underwater activities, should select their performances demonstrating their advantages. Although they can swim freely without trouble of umbilical cable, they are suffering from triple distress:
Although it is substantial to investigate and overcome this triple distress for realization of a sophisticated AUV, it is also necessary to construct practical and demonstrative vehicles based on the technologies in this decade despite of these handicaps. Considering advantages of AUVs, competitive AUVs at present, which should be constructed, can be classified into the following three species based on range and time of swimming, depth to dive, and complexity of work.
It is evident that the deeper the robot dives the more advantageous the performance is.
In addition to above three species,
for investigation of advanced AUVs should be included in the group of vehicles to be constructed. The fourth vehicle is generally designed to swim in a pool and seldom dives in the sea except demonstration of its high performance.
From Table 1, a list of AUVs (unmanned untethered submersibles) which have been constructed or planned in JAPAN, it can be said that the first species is quite dominant at present mainly because of simplicity of its mission. The same tendency can be observed for AUVs over the world.

2. VEHICLES OF UNIVERSITY OF TOKYO

A research group at Institute of Industrial Science of the University of Tokyo started research and development of AUVs which perform their missions perfectly independently from the operator. In other words, our vehicles should not depend on the command through the acoustic communication link to the mother ship. "PTEROA150" was constructed as a prototype vehicle of the cruising species in 1989. In 1992, "ALBAC" and "TWIN-BURGER" were constructed as a shuttle and a test-bed species, respectively. Although we have not a plan to construct a prototype vehicle of the third species, the TWIN-BURGER is anticipated to grow up in wisdom and performance as a matured third species.

2.1 PTEROA150 and R1

The PTEROA150 [1,2], which is shown in Fig. 1 and Table 2, succeeded for the first time in diving at the sea in November, 1990.
The vehicle has four echo sounders of 50 m range to detect the configuration of the sea bed and to keep a constant altitude. Figure 2 shows a trajectory of a sea trial which was conducted in December, 1992. At the first stage, the vehicle swam horizontally at 4 m depth for a predetermined time. Then, the vehicle descended till the bed was detected, and changed the control mode to the constant altitude swimming. In the figure, the configuration of the bed is drawn based on the data of echo sounders and it is shown that the vehicle succeeded in following the line of 10m altitude.
The objective of construction is to present an image of practical cruising type AUVs, to demonstrate their possibility, and to provide a model for basic research on technologies related to AUVs. The following are examples of effects of the construction of the PTEROA150. The PTEROA250 [1], which dives to the depth of 6,000m and swims for 2 hours over the bed, was designed by modifying the PTEROA150. The "Aqua Explorer 1000", constructed by KDD Co., Ltd. for survey of subsea cables in 1992, was designed on the same concept of PTEROA150's configuration. Though the PTEROA150 refuses the connection to the mother ship to be independent, the Aqua Explorer 1000 has a strong communication link to be remotely operated as a practical AUV. To develop a sophisticated intelligent control system, the adaptive controller system based on neural networks [3-5] is currently investigated using the model of the dynamics of the PTEROA150.
To emphasize the advantage of AUVs, long range swimming , i.e. long endurance, could be enhanced. Frequency of launching and recovery operations can be reduced resulting in decrease of probability of accident. For this purpose, it is necessary to reinforce the energy system, i.e. a dense, compact, reliable, and inexpensive energy source should be developed. When the capacity of energy system is large, we can adopt a system which has a fixed apparatus such as an engine and a generator. Considering these conditions, the University of Tokyo and Mitsui Engineering & Shipbuilding Co., Ltd. started from 1990 a joint project named "Development of Unmanned Untethered Submersibles for Survey of Mid-Ocean Ridges", in short, the "R1 Project" [6]. In the first stage of this project, we are constructing a prototype vehicle (cf. Figs. 3, 4 and Table 3) which dives to the depth of 400m for 24 hours. The energy system of 5kW output has been completed and installed in a corresponding pressure chamber.

2.2 ALBAC

Measurement of water column is also an important mission which can be done by AUVs of shuttle species. The system of the vehicle, which descends to a specified depth and returns to surface, can not be complicated because it is not necessary to pay continuously attention to obstacles and to analyze the data. The mission is so simple that the software is not necessary to have a complex structure to survive in a hostile environment. It can be said that the pop-up vehicle is a premature example of this kind. When the vehicle can move horizontally under the sea, it becomes more serviceable.
Distinctive features of the ALBAC [5] (cf. Fig. 5 and Table 4), which was constructed as a practical AUV for measurement of temperature of water column, are summarized as follows: 1) It consists of a cylindrical body, a pair of main wings, a tail plane, and a vertical stabilizing fin. 2) It has no propeller thruster but moves horizontally by gliding. It has not, therefore, a large capacity of battery cell and can provide a considerably large volume for pay load despite of its small dimensions. 3) It controls the pitching and rolling angles by changing the location of the center of gravity. Since the actuators are installed in the pressure chamber where the computers and interface boards are installed, the system becomes highly reliable without trouble related to connectors and cables.
In January 1993, the ALBAC carried out measurement of water temperature off Numazu, and demonstrated the capability of the shuttle species of AUVs.

2.3 TWIN-BURGER

In order to investigate more general and higher intelligent behaviors by AUVs, such as maintenance of deep sea station, drawing a detailed map around a specified point, cooperative works with several robots and divers, etc., we constructed a versatile test-bed vehicle named "TWIN-BURGER" in 1992 (cf. Fig. 6 and Table 5) as the third vehicle of the University of Tokyo [8], and are building his brother. Though the preceding vehicles were planned to be operated at sea and the depth rating is an important factor, the TWIN-BURGER as the fourth species was designed only for pool or shallow water tests considering that
Two pressure chambers, which are fitted at the top as shown in Fig. 6, are designed to provide sufficiently large dry volume for circuit boards and sensors. They can be easily opened for maintenance and reconfiguration without trouble of the wires and connectors. The TWIN-BURGER has two systems for broadcasting his idea to other agents, such that an ultrasonic transducers and five EL-panels at the head. It should be emphasized that a TV-camera is fitted at the front end of the lower cylinder for battery cells and a VTR. The image from this camera is processed by his computer and utilized for decision making.

3. Next Step

In order to realize competitive AUVs, it is necessary to carry out sea trials of the developed vehicles. But sea trials consume time and money, and also involve some risk of missing. Our vehicles seem willing to stay silently on the bottom. We cannot, therefore, frequently repeat sea trials. We should do basic research by computer simulation and by utilizing small test-bed vehicles.
While operating three vehicles and demonstrating their advantages, we are currently investigating the following issues in our laboratory:
Results of them will be converted to the actual vehicles, and the vehicles will get higher ability. This style of investigation can be done only when we have an actual vehicle which can be operated at the sea. Since construction of an actual vehicle is considerably expensive, we believe for making progress in AUVs that it is important to do a cooperative research work with the laboratory which has such a vehicle. Those who are interested in our vehicles are basically welcome to our laboratory.

References

  1. Ura, T.: "Development of AUV 'PTEROA', International Advanced Robotics, MBARI, (1990) pp.195-200
  2. Ura, T.: "Free Swimming Vehicle 'PTEROA' for Deep Sea Survey", Proc. ROV'89, (1989) pp.263-268
  3. Suto, T. et al: "Unsupervised Learning System for Vehicle Guidance Constructed with Neural Network", Proc. 8th. Intn. Symp. on Unmanned Untethered Submersible Techn., Durham, (1993) pp.222-230
  4. Ura, T. et al: "Identification of Motion of Underwater Robot with Neural Network", J. Soc. of Naval Architects of Japan, Vol.174, (1993) pp.887-892 (in Japanese)
  5. Yoshida, Y. et al: "Adaptive Neural Network Application to ROV Maneuvering", Proc. MTS'92,Washington.D.C, (1992), pp.1053-1059
  6. Obara, H. et al: "Basic Design of an Unmanned Untetherd Submersible with a Closed Cycle Diesel Engine", Proc. 8th. Intn. Symp. on Unmanned Untethered Submersible Techn., Durham, (1993) pp.411-419
  7. Kawaguchi, K. et al: "Development and Sea Trials of a Shuttle Type AUV "ALBAC"", Proc. 8th. Intn. Symp. on Unmanned Untethered Submersible Techn., Durham, (1993) pp.7-13
  8. Fujii, T. et al: "Development of a Versatile Test-Bed "Twin-Burger" toward Realization of Intelligent Behaviors of Autonomous Underwater Vehicles", Proc. OCEANS'93, Victoria, (1993) pp.1.186-1.192
Other References about Underwater Robots in URA Laboratory
Last modified: Tue May 16 18:21:35 1995

URA Laboratory, IIS, The Univ. of Tokyo / auvlab@iis.u-tokyo.ac.jp