The team has put together a specification sheet on the Ranger dexterous manipulator. Click on the picture below to load
the specification sheet.
Simulated Hubble Space Telescope Electronics Control Unit
task
23 May 2002
(update by Brian Roberts)
An attempt was made to perform portions of the removal and replacement of a Hubble Space Telescope (HST) electronics control unit (ECU).
Three ECUs are found on the Telescope as part of the rate gyro assembly. Each ECU enclosure measures 11" x 9" x 7.5" and the unit weighs 17.4
pounds. The ECU is secured inside HST by four keyway slot bolts and one connector drive mechanism. Removal of the ECU is accomplished by backing
off these bolts six turns, unfastening the connector drive mechanism on the bottom of the ECU, and maneuvering the ECU off of the keyway slot bolts.
The replacement ECU is then installed into its mounting position on the keyway bolts and the connector drive
mechanism is fastened. The four keyway bolts are then tightened to approximately 5 ft-lbf. EVA tether loops are provided for retention
during on-orbit change out. Two ECUs were replaced on HST during a visit by astronauts during the first servicing mission in December of 1993.
Below are pictures of the Ranger engineering arm performing parts of the ECU removal
task and an estimate of how long that part of the task would take. First, the arm (without an end effector) moves back to get the bare bolt drive
which is mounted on a tool post. Once the wrist docks with the bare bolt drive, the wrist rolls to remove the bare bolt drive from the tool post.
With the bare bolt drive attached to its wrist, the arm moves to one of the ECU keyway slot bolts. After turning the bolt, the arm moves away from the
bolt, moves to another bolt, turns it, and repeats the process for the two
remaining bolts. The arm then docks with a tool post which allows the
removal of the bare bolt drive. The wrist moves to another tool post,
"docks" with the right angle drive, and after the wrist rolls to
attach the right angle drive to the wrist, the arm moves to the connector drive
mechanism at the bottom of the ECU. While this is happening, a second arm
would move to a third tool post to remove the tether loop gripper. After
the "fingers" of the tether loop close around the tether loop on the ECU,
the right angle drive on the other arm drives the connector drive
mechanism which starts to lift the ECU out of its fixture.
A video of parts of this task appears
below under the May 10th update.
|
Task |
Time |
|
Attach bare bolt drive to arm |
2:00 |
Move arm (with bare bolt drive) to ECU |
0:55 |
Place bare bolt drive on lower left bolt |
0:51 |
Turn lower left bolt 6 times |
0:48 |
Move arm from lower left bolt to upper left bolt |
0:51 |
Turn upper left bolt 6 times |
0:48 |
|
Move arm from upper left bolt to upper right bolt |
1:13 |
Turn upper right bolt 6 times |
0:48 |
Move arm from upper right bolt to lower right bolt |
0:59 |
Turn lower right bolt 6 times |
0:48 |
Move arm from lower right bolt to tool post |
1:10 |
Attach right angle drive to arm |
2:00 |
|
(Attach tether loop gripper to other arm) |
2:00 |
Move arm (with right angle drive) to connector drive mechanism |
0:57 |
Close tether loop gripper on tether loop |
0:28 |
Place right angle drive on connector drive mechanism |
0:39 |
Unfasten connector drive mechanism to lift ECU |
0:20 |
(Open tether loop gripper from tether loop) |
(0:12) |
|
|
|
|
Total time |
17:35 |
2002 IEEE International Conference on Robotics and Automation presentations
15 May 2002
Two papers were presented at the
2002 IEEE International Conference on Robotics and Automation that was held in Washington,
DC the 11th through the 15th of May. The first paper, "Effects of Time Delay on Telerobotic Control of Neutral Buoyancy
Vehicles," (paper (1.2M
PDF)) was presented by Dr. Corde Lane. The paper was written by
Dr. Lane, Dr. Craig Carignan, Brook Sullivan, Dr. David Akin, Teresa Hunt, and Rob
Cohen. The presentation (presentation (3.9M
PDF)) included four video clips which appear below (click on the pictures to start the
videos).
Manipulation task - replacement box change out
(3.8M QuickTime)
This video clip shows two experienced test subjects
using Ranger Neutral Buoyancy Vehicle I to perform insertions and extractions of the
orbital replacement unit fluids box with 0 and 3 second time delay.
|
Peg and hole simulation
(1.1M QuickTime)
This video shows a subject successfully compensating for the largest fixed
error treatment. Although the fixed error can cause large deviations between the
commanded and actual displays, that deviation does not change. Therefore, the
subjects were able to learn how to compensate and complete the peg-in-hole task
with minimal extra effort, even with a three second time delay.
|
Simulation future work
(2.4M QuickTime)
The new graphical simulation uses higher fidelity texture mapping to produce a more realistic virtual environment. These simulations will allow testing with multiple arms. The smaller window in this video clip shows the simulated camera output from the video manipulator. Here the operator controls
both the video manipulator and the dexterous manipulator to extract a
mock-up of the Hubble Space Telescope electronics control unit. Towards the end of the clip, tracking mode is used by the
operator to allow the video arm to follow the dexterous arm.
|
Ranger II Operations
(764K QuickTime)
An experienced operator control the dexterous
manipulator using two three degree of freedom hand controllers. This
new 8-DOF manipulator will allow the study of performing tasks with time
delay.
|
The second paper, "A Skew-Axis Design for a 4-Joint Revolute
Wrist," (paper (1.2M PDF)) was written by Drs. Carignan and Howard. The
presentation (presentation (3.4M
PDF)) included five video clips that appear below (click on the pictures to start the
videos).
Inverse kinematics modes
(788K QuickTime)
|
Joint limit avoidance
(495K QuickTime)
|
Singularity avoidance (simulation)
(385K QuickTime)
|
Singularity avoidance (experiment)
(263K QuickTime)
|
This video clip exhibits roll, pitch, and yaw motion for two inverse
kinematics modes for the wrist. The first mode is 4-DOF control with
self-motion. It uses a generalized pseudoinverse method with a
self-motion component for avoiding singularities and joint limits. The
second mode uses 3-DOF control with independent hand roll. It uses an
extended jacobian technique to control tool tip orientation using the
first three joint axes and hand roll using a separate command.
|
The beginning of the clip shows the arm operating with
the 3-degree of freedom (DOF) wrist controller. The arm stops as it hits
its yaw limit. After the arm moves away, the 4-DOF wrist
controller is
turned on causing the pitch housing to be rolled out of the way (without
any input from the operator and without the tip of the tool
moving). The arm is no able to "push through" the
location where it hit the limit with the 3-DOF wrist controller.
|
These clips shows the behavior of the wrist as it nears a Type II
singularity where all four joint axes collapse into the wrist roll/pitch
plane. The tool axis starts out aligned with the pitch axis so that
already the wrist is down to 3 degrees of freedom (Type I singularity).
As the operator inputs a rotation about the tool axis, the yaw axis
starts rotating around to become aligned with the wrist roll/pitch plane,
but the wrist roll (forearm) is rotated away before the yaw axis can
reach the plane.
|
15 May 2002
Dr. Russ Howard presented a paper,
"Design of a Robotic Wrist and Tool-Exchange Mechanism for Satellite
Servicing," (paper (3.8M
PDF)) at the 36th Aerospace Mechanisms Symposium that was held in
Cleveland, Ohio May 14th through the 17th. The presentation (12.4M
PDF) included three video clips which appear below (click on the pictures to start the
videos).
Wrist joint motion
(556K QuickTime)
This clip shows the 4 degrees of
freedom of the wrist (wrist roll, wrist pitch, wrist yaw, hand roll)
being activated followed by the two tool drive (fast and slow).
The wrist camera and LED can also be seen.
|
Joint limit avoidance
(495K QuickTime)
The beginning of the clip shows the arm operating with
the 7-degree of freedom (DOF) controller. The arm stops as it hits
its yaw limit. After the arm moves away, the 8-DOF controller is
turned on causing the pitch housing to be rolled out of the way (without
any input from the operator and without the tip of the tool
moving). The arm is no able to "push through" the
location where it hit the limit with the 7-DOF controller.
|
End effector change out (bare bolt drive)
(443K QuickTime)
This clip shows
the arm (without and end effector) "docking" with the interchangeable end effector
mechanism, performing a hand roll maneuver, which allows the bare bolt drive to be removed from the tool post.
|
An attempt was made to perform portions of the removal and replacement of a Hubble Space Telescope (HST) electronics control unit (ECU).
Below is a video clip of the Ranger engineering arm performing parts of the ECU removal task. First the arm (without an end effector) moves back to get the bare bolt drive which is mounted on a tool post. Once the wrist "docks" with the bare bolt drive, the wrist rolls to remove the bare bolt drive from the tool post. With the bare bolt drive attached to its wrist, the arm moves to one of the ECU keyway slot bolts. After turning the bolt, the arm moves away from the bolt and places the bare bolt drive back on a tool post. Once the bare bolt drive is removed, the wrist docks with and retrieves the parallel jaw mechanism which has a set of "fingers" that fit around the tether loop. The "fingers" are closed around the tether loop as the video ends.
Below is a short video showing the engineering positioning leg, head, and dexterous arm being put together. The video is at five times the normal speed.
At least a couple of dozen people operated the arm. Some of the Ranger team, who have zero or very little experience operating robot arms, managed to lock on
to a bolt and undo it. It really is not too bad, once you have the right camera views.
We had one (and only one) observed uncommanded motion. It was a cartesian pitch motion
occurring while someone was applying various wrist orientation commands.
This occurred near the end of the day. No idea why... it was very brief, it undid itself, and never repeated.
You are cordially invited to an open house at the University of
Maryland Space Systems Laboratory Space Technology Development Center on
Wednesday, April 17th, from 3:00-5:00 pm. The occasion of this event is the
completion of the development unit for the Ranger Telerobotic Shuttle Experiment
dexterous manipulators. This robotic arm, with eight degrees of freedom and
interchangeable end effectors, represents a new state-of-the-art in space
flight qualified manipulators; it will be demonstrated in action
throughout the open house. Please come join us to celebrate this achievement, and
to see the components coming together for the four manipulators that
comprise the Ranger Neutral Buoyancy Vehicle II.
The SSL Space Technology Development Center is located on the fourth
floor of the University of Maryland Engineering Research Building (previously
the Myers Building) next to the College Park Airport. It is only a couple
of minutes away from the College Park campus, and easily accessible by car
or Metro. Refreshments will be served! Please feel free to contact me
for directions, or check the directions below. Also be sure to check out the Space Systems Laboratory web site
(http://www.ssl.umd.edu). I hope to see you there!