In general, space robots are more floppier (mechanically flexible) than
robots used on the ground. This is because it is expensive to
launch a lot of mass (weight) into space, and the microgravity
environment in space allows the joints of the arm to be much weaker
than a ground robot. As a result of these factors, space
tend to be long and comparatively floppy. A commonly known
example of this is the Space
Remote Manipulator (SRMS)
The above traits also means that it can be difficult to use the space
robot on the ground. This is true in the case of the robotic
systems for the HRSDM. Both the long Grapple Arm and the
Dexterous Robot arms cannot be used on earth. In order to
completely verify that our tools and procedures will work, we need to
either hang the robot with cables to reduce the effect of gravity, or
we need to simulate its motion with mathematical models of the robot
This second approach is what we intend to pursue to verify our tools
and procedures. I proposed to the HST project that we build a
robotic simulator that uses mathematical models to preduct the motion
of the flexible robot system, but use a strong industrial robot arm to
execute the motions. In this manner, we can pick up heavy
and instrument simulators without worrying about how to reduce the
effects of gravity. Since this system will be used to verify
tasks with contact dynamics, I have named this laboratory, the HRSDM
Contact Dynamics Facility (CDF).
This kind of approach to tools and task verification is similar to that
used by other space agencies such as the Canadian
Artist conception of the CDF. The Ground
Trainer from MDA
shown on the left with its system of cables to reduce the effects of
The CDF robot is on the right (yellow). In the
middle (green), is a model of the Hubble Aft Shroud, which simulates
the hardware on HST.
This image from Brentford Powell.
An example, of the industrial manipulator that we will use for the
CDF. This is a
that has a
payload capacity of about 300 pounds and a reach in excess
of 8 feet.
This image from here
The industrial robot will be controlled from a software simulation that
includes the dynamic models of the HST/HRV stack, the long Grapple Arm,
the two Dexterous Robots, and their joints. Here I am
the simulation system with their joystick controllers. On the
screen one can see the graphical representation of the motion of the
system. In the future, that motion will be sent to the robot
that it can move according to the simulation. A force-moment
sensor at the wrist of the industrial robot will feed forces back into
the simulation, closing the force-motion loop. So as I move
simulated robot around on the screen, so will the physical robot move
in the lab. This and the next three photographs are by Chris
In March 2005, during our PDR, Dr.
Riccardo Giacconi, a Nobel Laureate
visited our Robotics Lab
facility and operated the simulator.
In the fore ground, from left to right: Steve Queen (one of the
developers of the simulator), Dr. Giacconi, Dr.
Weiler (GSFC Director),
(Hubble Mission Manager - with hand waving), John
Lymer (Chief Engineer),
Jill Holz (Robotics Program Manager). In the background, we
see a model of the entire Ejection Module
with a model of the Dexterous Robot visible (white).
We asked Dr. Giacconi to have a try at the simulator. Here,
holding the Rotational Hand Controller (RHC). This commands
the robot end-effector in the rotational degrees-of-freedom.
the background from left to right: Jim Corbo (Lead Systems),
Frank Cepollina, Jill Holz, John Lymer, Carol Wong (Johnson Space
He quickly got the hang of it and was able to 'drive' the robot around
the model of the Hubble Space Telescope.
We concluded the Preliminary Design Review on March 25 2005, and both
the Goddard Review Team and the Independent Review Team agreed that we
were well prepared and ready for work towards our next major milestone,
which is the Critical Design Review in September. We were
encouraged by this, and hope that the new NASA Administrator will
restore funds in Fiscal Year 2006 in order to extend the life of HST.