Development of a Low-Cost Underwater Manipulator
by
Lauren Alise Cooney
SUBMITTED TO THE DEPARTMENT OF MECHANICAL ENGINEERING IN
PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF
BACHELOR OF SCIENCE
AT THE
MASSACHUSETTS INSTITUTE OF TECHNOLOGY
June 2006
© 2006 Lauren Alise Cooney. All rights reserved
The author hereby grants to MIT permission to reproduce and to distribute publicly paper
and electronic copies of this thesis document in whole or in part in any medium now
known or hereaftr created.
Signature
ofAuthor.
.... . _....................
.E
partment of Mechanical Engineering
May 18, 2006
Certified by ......................................................
-- . ...-
/'~
,,ssociate
,C
Accepted by ...............
~
....
John Leonard
Professor of Mechanical Engineering
Thesis Supervisor
...
- ........... ....,................
John H. Lienhard V
Professor of Mechanical Engineering
Chairman, Undergraduate Thesis Committee
MASSACHUSETTSINSTITUTE
OF TECHNOLOGY
AUG 0 2 2006
l
LIBRARIES
ARCH!WvE,,
2
Development of a Low-Cost Underwater Manipulator
by
Lauren Alise Cooney
Submitted to the Department of Mechanical Engineering
on May 18, 2006 in Partial Fulfillment of the
Requirements for the degree of Bachelor of Science in
Mechanical and Ocean Engineering
ABSTRACT
This thesis describes the design, modeling, manufacture, and testing of a low cost,
multiple degree-of-freedom underwater manipulator. Current underwater robotic arm
technologies are often expensive or limited in functionality. The goal of this research is
to produce a multiple degree-of-freedom manipulator utilizing relatively inexpensive,
commercial off-the-shelf servo motors. This project is designed for low-payload (< 0.5
kg) and shallow depth operation on a small remotely operated vehicle.
A completed underwater manipulator has been built using the new servo housing design.
Static and dynamic waterproofing techniques have proven satisfactory, offering a solid
design for waterproofing of servo motors. Preliminary tests of the integrated servo arm
system indicate that the arm will operate successfully in the underwater environment.
This design is anticipated to be used on an underwater vehicle in June 2006, as well as in
future undergraduate ocean engineering design subjects.
Thesis Supervisor: John Leonard
Title: Associate Professor of Mechanical Engineering
3
4
Acknowledgements
I would like to express my sincere appreciation to the following people that made this
project possible:
·
*
*
*
*
*
*
My advisor, Professor John Leonard.
The MIT ROV Team, in particular, Daniel Walker, Thaddeus Stefanov-Wagner,
and Kurt Stiehl. As well, thank you to the ROV Team's donors, who provided the
financial support for this project.
Mark Belanger and the Edgerton Student Machine Shop.
The MIT Towing Tank.
The Ocean Engineering Teaching Laboratory.
Zachary Gazak.
And finally, my family, for their love and support.
5
6
Table of Contents
ABSTRACT
3
Acknowledgements
5
List of Figures
List of Tables
8
9
Chapter
11
1 Introduction
1.1 Review of the State of the Art of Underwater Manipulators
11
1.2 Design Goals
13
1.3 Outline
15
17
Chapter 2 Mechanical Design
2.1 Motor Selection
2.2 Motor Housings
2.3 Linkages
17
19
20
2.4 Cost
2.5 Manipulator Arm Photographs
22
22
25
Chapter 3 System Modeling
3.1 Rigid Body Kinematics
3.2 Rigid Body Dynamics
25
26
3.3 Fluid Effects
3.3.1 Hydrostatics
3.3.2 Hydrodynamics
28
28
28
3.3.3 Previous Research on Hydrodynamic Modeling of
Underwater Manipulators
Chapter 4 Performance Assessment
29
31
4.1 Control Architecture
4.2 Servo Characterization
4.3 Arm Characterization
Chapter 5 Conclusions
31
32
32
33
33
5.1 Summary of Contributions
5.2 Future Research
Appendix A:: Mechanical Drawings
Bibliography
33
35
41
7
List of Figures
Figure 1.1: Schilling Robotics TITAN 4 [1]
Figure 1.2: SeaBotix TJG300 Three Jaw Grabber [2
12
12
Figure 1.3: MIT ROV Team 2004 Vehicle
14
Figure 1.4: Gripper Assembly and Gearing [5]
15
Figure 2.1: Arm with End-Effecter Defining Degrees-Of-Freedom
18
Figure 2.2: Spring-Loaded PTFE Seal [6]
20
Figure 2.3: Section View of Motor Housing
Figure 2.4: Photograph of Inside of Motor Housing
Figure 2.5: Photograph of Arm Assembly
Figure 2.6: Photograph of Arm in Extended Position
Figure 2.7: Photograph of Arm in Stowed Position
Figure 3.1: Reference Frames
Figure 4.1: Pulse Width Modulation
21
22
23
24
24
25
31
8
List of Tables
Table 1.1: State of the Art of Underwater Manipulator Technology
Table 2.1: Motor Torque Requirements
Table 2.2: Manufacturer's Motor Characteristics [5]
13
17
19
Table 3.1: Denavit-Hartenberg
26
Parameters
[8]
Table 4.1: Servo Characteristics
32
9
10
Chapter 1 Introduction
The ability to operate in the subsea environment presents many opportunities and benefits
in research, commercial, and military endeavors. However, the added challenge of
working underwater often results in technologies that are either very expensive or have
limited functionality, a trend largely reflected in underwater robotic arm technologies.
The goal of this research is to develop a working five degree-of-freedom (DOF)
underwater manipulator created from low cost, off-the-shelf parts.
1.1 Review of the State of the Art of Underwater Manipulators
Current underwater manipulator technologies largely fall into two categories: high
functionality at high cost, or limited functionality at low cost. Robotic arms with
multiple degrees of freedom are typically designed for medium- to large-scale remotely
operated vehicles (ROV's). These manipulators are often expensive, heavy, and large.
For example, the Schilling Robotics TITAN 4, shown in Figure 1.1, is a popular deep sea
manipulator employed on large ROV's working in the offshore industry. This arm is
made primarily out of titanium and is powered by hydraulics. Although the TITAN's
seven degrees-of-freedom are an attractive feature, due to its large weight and expense
(76 kg in seawater, >$150,000), this class of manipulator is not feasible for smaller lowcost vehicles.
At the other end of the spectrum, lower cost manipulators commonly have much less
functionality. The SeabotixTJG300 Three Jaw Grabber, as shown in Figure 1.2 and as
described in Table 1.1, has only a gripping mechanism. Because of this limited
performance, a greater dependence is placed on vehicle maneuverability in order to
achieve complex manipulator tasks.
11
Figure 1.1: Schilling Robotics TITAN 4 []
Figure 1.2: SeaBotix TJG300 Three Jaw Grabber
12
[2]
Table 1.1: State of the Art of Underwater Manipulator Technology
SeaBotix, Inc. [2] Inuktun Servies
Ltd. [3]
Schilling
Robotics
International
Submarine
Systems, Inc. [4]
Engineering
Ltd.
[4]
Model
TJG300
Micro
Manipulator
Titan III s
ISE 7F
DOF
Gripper
Gripper
6 plus gripper
6 plus gripper
Actuator
Lead Screw
N/A
Drive Mechanism
Hydraulic
Cylinders
Hydraulic
Cylinders
Jaw Closing
Force
14 lbs
5 lbs.
1000 lbf
330 lbs
Price
$1,995
$2,700
$149,500
$70,000$250,000
1.2 Design Goals
The design of an underwater manipulator encounters challenges from the combined
complexities of subsea operations and robotic arm design, which are further intensified
when trying to do so in a cost-efficient manner. The waterproofing of electronics and
other sensitive elements is critical to a working system. System design is driven by
desired payload capacity and end-effecter positioning and speed, as well as additional
forces from static and dynamic fluid effects.
Design considerations for this arm are made with the anticipation of functioning on a
small remotely operated vehicle (12 x 12 x 12in) built by the MIT ROV Team to compete
in the national ROV competition organized by the Marine Advanced Technology
Education Center (MATE) and the Marine Technology Society (MTS) ROV Committee.
The 2006 MATE ROV Competition is to be held at the Neutral Buoyancy Laboratory at
the NASA Johnson Space Center in Houston, Texas. The competition takes place in a
fresh water pool, where the maximum operating depth is 40 feet. There are several tasks
that the ROV must complete within the 30 minute competition time limit. The ROV
must transport an electronics module weighing 1.5 kg underwater to the pool bottom, and
secure this module into a frame. The module door needs to be opened-requiring 1 N of
force-so that a 1" PVC probe weighing 0.3 kg underwater may be inserted into the port
on the module.
13
In addition, there are design requirements created by the ROV itself. Low weight and
small size of the stowed arm are essential in minimizing manipulator effects on vehicle
operations. The arm should also be capable of operating off of a bottomside power
supply rated at a nominal 12 V and 20 Ah. Topside communication with the arm will be
through a fiber-optic tether. Of overarching concern for the arm is budget. The ROV
Team operates off of donations from both industry and academia. Keeping the arm costeffective (<$500) would make it a feasible enterprise, as well as increase the viability of
reproduction.
Figure 1.3 shows the ROV Team's vehicle for the 2004 competition. The gripper
assembly in Figure 1.4 was used on this vehicle to retrieve a weighted 2" PVC pipe off of
the bottom of a swimming pool and to bring it up to the surface. Although this gripper
does not have much in the range of maneuverability, it was acceptable for this mission, as
the only functional requirements were the ability of the ROV to land within grasping
range, and that the gripper could maintain a tight hold on the object.
Figure 1.3: MIT ROV Team 2004 Vehicle [5]
14
Figure 1.4: Gripper Assembly and Gearing [5]
For this year's competition, however, relying on the ROV's propulsion/control system to
perform tasks such as plug the instrumentation probe into the module port could be time
inefficient and difficult. An underwater manipulator with high functionality would
provide several advantages for the ROV competition, and to underwater operations in
general. A multiple degree-of-freedom robotic arm would allow for more precise control
in manipulation tasks, as opposed to a vehicle with a single gripper which would require
high levels of hovering control.
1.3 Outline
This thesis documents the development of a low-cost underwater manipulator. The
mechanical design process is described, including motor selection, housing design,
manufacturing, and cost. Modeling of the system is then performed, for rigid body
kinematics and dynamics, and an evaluation of fluid effects. Preliminary performance
assessment, conclusions and suggestions for future research are discussed.
15
16
Chapter 2 Mechanical Design
This chapter describes the mechanical design and manufacturing of the manipulator.
Component selection, development, and integration are discussed. Photographs of parts
and the assembled arm are shown in Figures 2.4 through 2.7. A bill of materials is listed
to illustrate the low-cost aspect of the design. Mechanical drawings for components can
be found in Appendix A.
The manipulator is made up of five direct serial drive revolute joints. The end-effecter
was not developed for the purposes of this thesis. The configuration of the manipulator is
similar to that of a human arm, with a two degree-of-freedom shoulder connected to the
base (01 and 0 ), a one degree-of-freedom elbow (03), and a two degree-of-freedom wrist
(04 and 05) (Figure 2.1).
2.1 Motor Selection
Hobby servos are an appealing technology, as they are readily available with gear trains
and feedback positioning electronics at a relatively low cost. A new breed of robotic
servos marks a departure from standard model airplane servos, as they boast high torque
and over 180 degrees of rotation.
Torque requirements for each motor are determined using a static moment balance, where
buoyancy effects are considered and hydrodynamic effects are small at very low speeds.
These calculations include the weight of an end-effecter and a 1 N weight. Table 2.1
presents the torque requirements for the case of hydrostatically optimized motor housings
that achieve neutral buoyancy.
Table 2.1: Motor Torque Requirements
Torque Requirement
In Air
In Water
Motor 2
4.53 N-m
1.309 N-m
Motor 3
1.856 N-m
0.562 N-m
Motor 4
0.473 N-m
0.158 N-m
17
!, ___' ':··r:
.
~ ~ ~ ~~~~··
..........~
,
",
t;
NJV
03
0f1
0 f Z6
04
,L.
-1H
Figure 2.1: Arm with End-Effecter Defining Degrees-Of-Freedom
18
Two servo motors produced by Hitec, Inc, make up the drive framework for the
manipulator: the HS-805BB and the HS-755HB. The HS-805BB is used to drive the
arm's shoulder and elbow joints, as these joints require more torque, but can also handle
a larger size motor as the moment arm created about the base motor is minimal. The
HS-755HB motor drives the two wrist joints as they do not require as much load carrying
capacity.
Table 2.2: Manufacturer's Motor Characteristics [6]
HS-755HB
HS-805BB
Stall Torque
13.2 kg*cm
19.8 kg*cm
Speed (no load)
0.23 sec/60 deg
0.14 sec/60 deg
Operating Angle
±90 deg
±90 deg
Operating Voltage
4.8 - 6.0 V
4.8 - 6.0 V
Current Drain (no
285 mA
830 mA
Dimensions
59 x 29 x 50 mm
66 x 30 x 58 mm
Weight
110g
152 g
Price
27.99 USD
39.99 USD
2.2 Motor Housings
The motor housings are constructed out of black nylon 6/6, chosen for its good tensile
strength, dimensional stability, machinability and relatively low cost compared to plastics
of similar performance level. Large pieces of stock, however, were more expensive than
several smaller pieces of stock with the same cumulative volume. Therefore, the housing
is made up of two smaller equal sized blocks. A standard face seal O-ring groove creates
the static seal between the two housing components and the outside environment.
The dynamic seal for the motor shaft is provided by a spring-loaded PTFE seal (Fig 2.2).
These seals employ U-Cup lip seal geometry using a canted coil spring that creates a
sealing force that is also energized under dynamic conditions. The ratings for these seals
are well within the operational limits for this project (pressure x velocity limit of 75,000
psi-fpm). For deeper applications, an alternative dynamic seal would need to be applied.
19
Figure 2.2: Spring-Loaded
PI1E
Seal
71
L'J
Shaft seals prove difficult for hobby servos as a servo horn is directly connected to the
final gear stage, leaving no shaft exposed to create a seal around. The final design
consists of a commercially available inline shaft adaptor connected to the servo horn,
around which the shaft seal is created.
Plastic flanged sleeve bearings are utilized at both the output shaft and linkage pivots.
The bearings chosen are of material Rulon LR, preferred for its low friction and because
it does not absorb water. These bearings act as constraints in the radial direction and also
handle radial and thrust loads. The bearings used at the linkage pivots also serve as a low
friction pivot point.
2.3 Linkages
Linkages are constructed out of 3/8" clear polycarbonate, a material that exhibits good
strength and relative ease of machining, yet low weight as compared to metals such as
aluminum. Each motor driven linkage operates in parallel with a load-sharing linkage
connected to the motor housing via a freely rotating pivot. Linkages are secured to shafts
using a dowel pin.
20
Motor Outpu
Ro~rirn
I I
I I ~"
Spring-Load
Spring-Load
Shaft Attach
Servo Horn
servo
O-Ring
Motor Housii
I
./
~~~~~~~~~
Bearing
Linkage Pivot Rod
I~~~~~~~l
lr
Figure 2.3: Section View of Motor Housing
21
2.4 Cost
Table 2.3 presents the cost of materials used in the final manipulator design, not
including the cost of an end-effecter or PIC microcontroller for control. As can be seen,
the total cost is significantly less as compared to even the single degree-of-freedom
grippers listed in Table 1.1.
Table 2.3: Bill of Materials
Unit Price Quantity Total Price
IMotors
IHitec HS-755HB 1/4 Scale Karbonite Gear Servo U
Hitec HS-805BB Mega 1/4 Scale 2BB Servo U
MaterialsBlack
Nylon 6/6
Clear Polycarbonate Sheet
Other
Aluminum Shaft Attachment
Sleeve Bearing
Spring-Loaded PTFE Seal PTFE, 1/16" Width, 1/4" Shaft Dia, 3/8" Seal Od
Unshielded 22AWG UL2464 3 COND, 100"
|
$27.99 |
$39.99
21
$55.98
3 $119.97
$114.80
$23.79
1
$114.80
$23.79
$9.95
$1.88
$8.00
$31.75
5
10
5
1
$49.75
$18.80
$40.00
$31.75
Grand Total:
$454.84
2.5 Manipulator Arm Photographs
MOIX1011-4Aft
'.
i' ,
Figure 2.4: Photograph of Inside of Motor Housing
22
.
. j, -
, S ,, i z .., 4
-
.
10
:
T....
M.
: I
I
-:17
.1.111
1.11111%
link--.
Figure 2.5: Photograph of Arm Assembly
23
I .- ..,
Figure 2.6: Photograph ot Arm in Extendaea osition
Figure 2.7: Photograph o0 Arm in Stowea Posltlon
24
Chapter 3 System Modeling
This chapter presents the kinematic and dynamic analysis of the manipulator. In addition,
fluid effects are defined and discussed.
3.1 Rigid Body Kinematics
The system in consideration can be idealized as 5 revolute joints, each rotated at an angle
0i with respect to the orientation of its connecting serial linkage of length fi (Fig 3.1).
The Forward Kinematic Equations are used to describe the location and orientation of the
end-effecter in the fixed vehicle reference frame 0- x 0 yoz 0 . For the purposes of this
project, the vehicle is considered to be stationary within the fixed earth reference frame.
The position of the end-effecter is described by the three coordinates [Xe,ye, Ze].
I
r;
-11
04
03
t2
05
t3
t4
Figure 3.1: Reference Frames
Forward Kinematic Equations:
Xe = {2 cos82
'e = { 2
Ze
=
*1
cos 0
+
e3 cos(02 +
3
)+ f4 cos(0
3 cos( 2 + 0 3 )+ e4 cos(
2
2
+ 0 4 ) + (w4
+
5 )sin(02 +03 +
+03 + 04 ) + (w 4
+
5 )sin(0 2
+3
+f2 sin0 2 + 3 sin(02 +0 3 )+ e4 sin(0 2 +03 +0 4 )-(w
25
4
4
)}cos01
+ 03 + 04 )}sin0 1
+/I 5 )cos(0 2 +03 +04)
Table 3.1: Denavit-Hartenberg Parameters
i
ai
ai
di
i
°
e'
a0
0
0
02
t3
0
0
03
4
W4
-90 °
'4
04
5
i5
0
0
05
1
O
-90
2
E2
3
8]
where:
ai
ai
d
0i
Link length, distance from Z i to Zi+ along Xi;
Link twist, angle from Zi to Zi+, about Xi,
Link offset, distance from Xi_, to Xi along Zi, and
Joint angle, angle between Xi-, to Xi along Zi.
Homogeneous transformation Ai:
Soi Cai
Soi
o
0
Soi a ai
CO
i Ca i
-Ci S a i
Sai
Cai
0
0
co
i
aisOi
1
3.2 Rigid Body Dynamics
Characterizing the motion of the arm within the fixed vehicle reference frame is critical
to understanding arm dynamics. These equations can become quite complicated. For the
purposes of this research, primary interest is with regard to large motions of the arm.
Therefore, in the analysis in this section, rotations in 01, 02 and 03 only will be considered,
and the joints at 04 and 05 will be rigid.
The Jacobian matrix J describes the relationship between joint angular velocity
, i and the velocity of the end-effecter in the reference frame:
26
r
CI
o
Cc
+)
0+
C
'O5
+
+
CD
L.
0
+ )
:,
+
CD
r
+O
"
+
C+'D
+~ICD
+
+'
0L
+
2
+
(CN
r,
+
C
+
(_
r
. Co
-I-
Co
r0
+
CN
+
+>
C'
+ +
+
C4
0
+
C-
0
oCD CO
'
+_
-
Co
4D
1
a',
+
<-
("
e1
@4
+
+
r
C
o
C)
C @,
C--
c)
C")
Co
+
+
C
I
C-
CO
04
0
K'
I
tQ
~C'I
I
I
a'@
'
a'
04
o
C)
(NJ
C@
(8Q
0 ( 0 ( (Z-Q
c c( cO
'IQ
( ( ( 0 tCZ0 (--
III
I
8t)s
ZmI
CO
(C
r
*q
o
C'D
-I
(
IIcC;
~ rg~OI
CO
cO
II
II
--.
C'q
+-N
,
+0a
+
o
o
o
+
I
('14
II
C)
+
+
(f)
o
0
.C
3.3 Fluid Effects
Underwater operation adds both static and dynamic complexity. The hydrodynamic
modeling of underwater manipulators is quite complex, therefore, discussion in this
section is aimed more at an overview of the hydrodynamic effects and their possible
impact on performance, rather than computation of exact values.
3.3.1 Hydrostatics
In static analysis of underwater bodies, both the gravitational force acting on the body
mass and buoyancy force must be considered. Archimedes's principle states that the
buoyant force is equal to the weight of the fluid displaced. Therefore,
Fbuoyancy = Pwater
submerged
g
Fgravity= mbodyg
FtotaI = Fbuoyancy-Fgravity
3.3.2 Hydrodynamics
Added mass is the effect of fluid inertia in the environment on a moving body. This is
not a phenomena that extends to above water operations, however, the effect is vastly
more significant subsea as the density of water is much more comparable to the density
of the body of interest. The added mass coefficient is dependent on body geometry and
motion.
Fluid viscosity also creates drag and lift forces on a body. Drag acts in parallel with the
flow velocity on the body:
Fdrag = pU 2 A Cd (Re)
where p is the fluid density, U is the flow speed, A is the projection of the body in the
direction of flow, and Cd is the drag coefficient which is dependent on Reynold's
number:
Re = plUID
'U
where gt is the dynamic viscosity of the fluid. The lift force acts normal to flow velocity:
Flif 1= pU2ACI (Re,a)
where the lift coefficient C1 is a function of Reynold's number and a angle of attack.
28
For a cylinder in non-uniform flow, the total added mass and drag forces can be found
using Morison's equation, which relates the inertial and drag forces:
dF =-C, * pAUdl -C * pWIUIUdl
3.3.3 Previous Research on Hydrodynamic Modeling of Underwater Manipulators
The complexity of hydrodynamic modeling of underwater manipulators is largely due to
a lack of understanding of the 3-dimensional hydrodynamic effects of manipulator
systems. To be considered are flow effects around rotating joints and linkages in close
proximity, and their dependency on rapid movement and unsteady motions.
It has been determined through experimental research that using constant added mass and
drag coefficients for a manipulator body are not sufficient for complete analysis [9]
Leabourne and Rock [10]expanded on this research, and found the drag and added mass
coefficients to be dependent on elbow angle in their two degree-of-freedom arm model.
In these tests, the maximum value of the drag coefficient for the manipulator model was
found to be 3.
29
30
Chapter 4 Performance Assessment
This chapter describes the performance of the arm design presented in previous chapters.
The overall arm control architecture is discussed, and preliminary testing of motors and
the arm are assessed.
4.1 Control Architecture
Hobby servos are controlled using pulse width modulation (PWM). The servo angle is
regulated by varying the signal pulse width, where 1.0, 1.5, and 2.0 msec prescribe motor
rotation angles of -90 deg, 0 deg, and +90 deg respectively.
Signal
50 Hz Period
1 .0 - 2.0 msec PulseWidth
*
-
Tine
Figure 4.1: Pulse Width Modulation
The control system and user interface for the manipulator were developed by members of
the ROV Team. The arm is operated using a master-slave system. The master is topside,
and is of identical kinematic design to the actual manipulator. Each joint of the master
has a position potentiometer that measures the motion of the operator as the master is
moved to simulate the position of the actual arm. The desired position is chosen in
parallel with real-time video feed of the manipulator's location.
The control system employs two PIC microcontrollers operating in closed loop, one-way
communications control. The topside PIC reads in the potentiometers in the joints of the
master arm, converts the values into motor angles and sends the info down a fiber optic
tether. The PIC on the bottom side receives the data, converts the motor angle and sends
the signal to the motors.
31
4.2 Servo Characterization
An HS-805 servo was tested using a PIC outputting PWM. The motor was enclosed in
its waterproof housing for both air and water tests. Maximum slew rate was determined
by measuring the time for the motor shaft to rotate clockwise, and averaging this rate
over 10 trials. Power consumption without load was measured as the current draw
observed under no load, with a voltage limited power supply at 6.0 V. Maximum power
consumption with load was the largest current draw observed over the testing period,
under voltage limited power supply of 6.0 V. These characteristics are shown below in
Table 4.1. As can be seen, the max slew rate and power consumption without load are
fairly close to those expected by the manufacturer's specifications. The higher measured
values are expected, as the motor shaft, even under no load, is experiencing additional
torque due to the spring-seal. These discrepancies are not large enough to be cause for
concern with regards to housing and shaft seal effects on motor performance.
Table 4.1: HS-805 Servo Characteristics
Manufacturer's
Specifications
Motor in Air
Motor in Water
Max Slew Rate
180 Deg / 0.42 Sec
180 Deg / 0.54 Sec
180 Deg / 0.57 Sec
Power Consumption
w/o Load
4.98 Watts
5.64 Watts
5.88 Watts
Max Power
Consumption w/
Load
N/A
11.22 Watts
12.66 Watts
4.3 Arm Characterization
The tests performed with the assembled arm consisted of controlling only the first and
second motors, one at a time. The arm was assembled using the first four motor housings
and the two linkage pairs as seen in Figure 2.5. The motor that was being tested was
given a PWM signal to control the position of the motor shaft. Due to time constraints
and technical difficulties, the abilities of these motors to provide sufficient torque was
only observed qualitatively. Both motors were able to drive their successive linkages and
attached motors without stalling. The slew rate did not appear to vary significantly from
tests with no load.
32
Chapter 5 Conclusions
This chapter concludes the thesis by summarizing the contributions of this research and
making recommendations for future work.
5.1 Summary of Contributions
This thesis presents the foundation for a low-cost, several degree-of-freedom underwater
manipulator for a small ROV. The primary achievement is the design of a waterproof
servo motor housing consisting of static and dynamic seals. Several of these housings
were manufactured and proven to be watertight, and to have minimal effect on motor
performance. Kinematic and dynamic analysis was performed, as well as the discussion
of fluid effects for the manipulator. In addition, a preliminary manipulator arm was
assembled, which required the design and manufacturing of motor shafts, linkages, and
supports. Preliminary underwater tests of the motor housings and of the assembled
manipulator were successful, and encourage the continued development of the arm.
5.2 Future Research
In order to make this arm ready for field operation, the motor housings should be tested
to maximum designed depth.
If more extreme operating depths are desired, a new
pressure compensation system must be developed, such as the oil regulating bladder
system described by Di Pietro for the ROV JASON manipulator [11]. An end-effecter is
currently being developed for the arm. Continued arm characterization needs to be
performed using the end-effecter with various payloads.
The primary drawbacks to the current motor housing design are cost of material and
manufacturing time. A large portion of the nylon stock used to create the housing is
hollowed out and discarded during manufacturing, which is an inefficient use of time and
money. Using a readily available, inexpensive, waterproof housing would be a
significant improvement on manufacturing and cost. Another interesting avenue would
employ inexpensive DC motors with positioning electronics. DC motors would reduce
complexity in the shaft seal and possibly result in more compact motor housings.
33
34
Appendix A
0.400
0.400
--- -
I-
Cbs
--
7-I
1
c)
Is
++- ++
tco
,q
COI
',0~
J_
._
.
A
C
-Q
co
KtJ_-
J
II
I
__
.-
3.080
2.818
I
I!1
I
l
[H4.
.-
t:O
A,----
d
I.-
If
A----
/er Link
7.548
Link
9.500
~9.500
;
35
a)
0
Eo
an
2)
oo
IX
1)
('
O O
-1 607'0
I--
gL'L L
OOL0'O
9L0'0
S
+ -c
oo
I'
I- D
A
..
--
II -1
L
I
.
36
0a
._
c"
0
0CO
c
-------- 1
--------I
--------
U
cO
cO
0f
--------------_I
'I-,
(_X_!
X - ,.
,,--..
o
ol
\
og'o
Jii
2C9, 1
--
L176'
37
&
t
-- jl
-
+,
.~1
I
'0
0
co
o
11-
--------
-
|
-
a.
0
a,)
Oo
C
=D
08'L
,J
-
------1---
----------l
--
--
- - - -
-
11 -
E
n
0
o
m
-
L-C
O -CT) C
I D
t79t7'L
-. E3-
060'L
1
-
F--HI
· t
I
I
I
Lil
I=
L_
L I
.- l
I
Ot77n
7YVV
099'L
--I
I-
-------
:p
t-
H_
-F
2l
i
i-
- =
I
r _
-----
3
I
-IL
I -I
.
_BS
O Cl1
YO
'
VY6 L
_~
E
I
I
39
40
Bibliography
[1]
"Manipulator Comparison." Product Literature. Schilling Sub-Atlantic. 15Apr.
2006 <http://suab4599.demonweb.co.uk/manipulators.html>.
[2]
"LBV Grabber." Product Literature. SeaBotix, Inc. 15Apr. 2006
<http://www.seabotix.com/products/pdf_files/LBV_GrabberA4.pdf>.
[3]
"ISL MicroManip." Product Literature. RoboProbeTecnologies. Inc. 15Apr. 2006
<http://www.roboprobe.com/DOWNLOADS/D 13_IK_MICROMANIP.PDF>.
[4]
Yuh, J, and M West. "Underwater Robotics." Advanced Robotics 15 (2001): 609639.
<http://www.ingentaconnect.com/search/article?title=underwater+robotics&titlet
ype=tka&year_from= 1998&yearto=2005&database=
1&pageSize=20&index=5>
[5]
Stanway, M. et al. "ROV Castor: Design of a Multi-Mission Remotely Operated
Vehicle." 2004.
[6]
"Giant Scale Servos." Product Literature. Hitec RCD USA Inc. 03 Feb. 2006
<http://www.hitecrcd.com/homepage/product_fs.htm>.
[7]
"FlexiSealTM." Product Literature. Parker Seals. 10 Mar. 2006
<http://www.parker.com/packing/cat/english/5315_FlexiSeal.pdf.
[8]
Craig, John J. Introduction to Robotics. 2nd ed. Reading: Addison-Wesley
Company, Inc, 1989.
[9]
McLain, T., Rock, S., and Lee, M. "Experiments in the Coordinated Control of
an Underwater Arm/Vehicle System." Journal of Autonomous Robots 3 (1996):
213-232.
<http://www.ingentaconnect.com/content/klu/auro/1996/00000003/00000002/001
16327>
[10]
Loeabourne, K., and Rock, S. "Model Development Of An Underwater
Manipulator For Coordinated Arm-Vehicle Control." (1996). 15 Apr. 2006
<http://sun-valley.stanford.edu/papers/LeaboumeR: 98.pdf>.
[11]
Di Pietro, David. Development of an Actively Compliant Underwater
Manipulator. Diss. Massachusetts Institute of Technology, 1988. 20 Jan. 2006
<http://dspace.mit.edu/handle/1721.1/14582>.
41
MITLibraries
Document Services
Room 14-0551
77 Massachusetts Avenue
Cambridge, MA 02139
Ph: 617.253.5668 Fax: 617.253.1690
Email: docs@mit.edu
http: //libraries. mit.edu/docs
DISCLAIMER OF QUALITY
Due to the condition of the original material, there are unavoidable
flaws in this reproduction. We have made every effort possible to
provide you with the best copy available. If you are dissatisfied with
this product and find it unusable, please contact Document Services as
soon as possible.
Thank you.
Some pages in the original document contain color
pictures or graphics that will not scan or reproduce well.