The target of research is autonomy in execution of dextrous tasks using manipulators on a free swimming underwater robot . The aim is to constructing coordinated control system of multiple manipulators that is capable of autonomous position and force control of end-effector of mission arm precisely.
The present work deals with position control of
end-effector of mission arm using multiple stabilizing arms to
counteract reaction forces and moments at the mission arm, and
disturbances to the vehicle.
(1)N.Kato,D.M.Lane,"Coordinated Control of Multiple
Manipulators in Underwater Roots", 9th Int. Symp. on
Unmaned, Untethered Submersible Technology, pp.34-50,1995
(2)N.Kato,D.M.Lane,"Coordinated Control of Multiple
Manipulators in Underwater Roots", 1996 IEEE International
Conference on Robotics and Automation, pp.2505-2510,1996
Stabilization of motion of model of an underwater robot equipped with a manipulator with assistance of the pair of 2-motor-driven mechanical pectoral fin
The experimental results in horizontal plane on the stabilization of motion of model of an underwater robot equipped with a manipulator with assistance of the pair of 2MDMPFs instead of the paddling arms used in the simulation are given here.
A. Model of Underwater Robot with a Manipulator
Fig.1 shows the model of the underwater robot with a manipulator. The dimensions of the model are as follows: Length*width*height = 0.9m*0.6m*0.5m. The mass of the model is 25.95 Kg. The manipulator consists of 3 links where 3 DC motors and 3 potentiometers are installed. The dimensions of the manipulator are as follows: Length width *height = 0.651m*0.040m*0.0.095m. The distance between each link is 0.2 m. The mass of the manipulator is 2.55 (kg). The pair of 2MDMPFs were installed at 0.34 m behind from the root of the manipulator.
Fig.1 Model of Underwater Robot with a Manipulator
B. Setup of Experiment
The model of an underwater model and the carriage for the measurement of the motion in the horizontal plane was rigidly connected by a support rod which is located at 0.34 m behind from the root of the manipulator. The fish robot was immersed in water at half depth of the water circulating tank
C Control Algorithm
Fuzzy algorithm was used for the control of yaw motion of the model of an underwater robot with a manipulator. The control variables were a pair of phase differences between the lead-lag motion and the feathering motion on both sides. The amplitudes of both motions were fixed.
D Experimental Results
Fig.2 shows an example of experimental results of yaw motion of the model during a period (11.1 s) of oscillatory motion of manipulator. The position and the posture of the model with a manipulator is illustrated each 0.6 s.We find that the strong interaction of motion occurs between the model and the manipulator.
Fig.3 shows the ratio of root-mean-square of yaw motion of the model without control to with control during 5 periods of sinusoidal motion of the manipulator against the variables of period of manipulator from 5.5 s to 11 s and that of pectoral fin from 1.0 s to 2.0 s. We can see that slower motion of manipulator produces better performance of control that is hardly dependent of period of motion of fin, and that the control performance increases as the fin moves faster with faster motion of the manipulator.
Fig.2 Motion of Model of an Underwater Robot induced by an Oscillatory Motion of a Manipulator
Fig.3 Ratio of Yaw Motion of Model with Control to without Contro
(1)N.Kato: Application of swimming functions of aquatic animals to autonomous
underwater vehicles, Proc. of OCEANS'99 (MTS/IEEE), 1999
(2)N. Kato ,"Pectoral
Fin Controllers in “Neurotechnology for Biomimetic Robots,”
edited by J. Ayers, J.J. Davis and A. Rudolph, The MIT Press, pp.325-347,2002
This research deals with a development of an automatic restoration devive
of a capsized small marine vehicle by use of Eskimo-roll control. Eskimo-roll
is usually used to resore a capsized kayak by using a paddle. Eskimo-roll
is composed of long-roll, which is performed in a vertical plane, and short-roll,
which is performed in three-dimensional space.
We carried out the study of long-roll by simulation and experiment. We
performed the motion analysis of a human doing the long-roll and constructed
the dynamic model using coordinated control of maultiple manipulators. We found that the agreement between the simualated results and the measured
results are fairly well.
Next, we developed an automatic restoration device composed of a manipulator
with 3 links as shown in the following figure. We were successful in resoring
a capsized kayak model by the device.
