CMU Robot Inventory
This list was put together in 2016 by Chris Atkeson.
Clicking on images will get you a higher resolution image, if it is
available.
Previous Inventories
Robots at the Field Robotics Center
Robots built at CMU
On display at CMU: Newell Simon Hall (NSH end of bridge to Wean Hall).
CMU Direct Drive Arm II
On display at CMU: Newell Simon Hall (NSH end of bridge to Wean Hall).
Scientific American.com: Radioactive Robot: The Machines That Cleaned Up Three Mile Island
Remote Reconnaissance Vehicle (1983, Whittaker):
TMI 1, left at site.
TMI 2: Pennsylvania's State Museum in Harrisburg.
Workhorse robot on display in NREC lobby.
TMI and Fukushima: How Robots Were Used in Response
Terregator (1984-1991, Whittaker)
From: James Montgomery
We have Terregator on display here at NREC in our lobby
The Terregator Mobile Robot
Lee Champeny-Bares, Syd Coppersmith and Kevin Dowling
CMU-RI-TR-93-03
Navlab series
Navlab 1 (1986) was actively working as the
power source for farm equipment that feeds animals from a grain
silo. Some farm in Zelienople that a farmer had given Omead permission
and the ability to fly and test his helicopters.
Unfortunately the Navlab5 minivan was owned by Delco/Delphi. A couple
years after No Hands Across America they asked us to return it and from
what I hear, they had it crushed. :-(.
Pluto: Moravec's 6-DOF wheeled system.
The last Pluto existed until 9 years ago, when it was dumpstered
during a purge. (Red)
Donated to the Boston Computer Museum, which closed.
Moved to
Computer History Museum
Neptune (1984): simple three-wheeled system, became the workhorse for
Chuck Thorpe, Larry Mathies, Alberto Elfes, Rich Wallace
Neptune (and two darker blue copies made by other labs) were in the
catacombs under the highbay as recently as a year ago (2016).
Damaged in Newell Simon Hall flooding, and thrown out.
Uranus: (1985, Greg Podnar)
Mecanum Wheels
It's still at Seegrid. [Moravec]
Ambler
at NREC
Tesselator
Post Gazette article
Location?
Dante I
Cannibalized to make Dante II.
Dante II
Article
Dante II is at NREC.
Leg lab
We need to get the early CMU ones back at CMU.
Spinner - NREC
Highlander - Robot City
Sandstorm - Robot City
Boss - NREC
HERB - NSH - Cannibalized
Chimp - NREC
From: Howie Choset
I would be more than happy to donate old dead snake robots hanging
on the wall in my lab.
From: Manuela Veloso
I have tons of historical data, including videos and the earlier
physical robots, of all our robot soccer, and also the AIBO, the NAOs,
and our CoBot robots...
From: Geoff Gordon
In my office I have a Nursebot project robot -- I'd be happy to
donate it to the cause if it is helpful. (Right now its main use
is as an objet d'art for my office.)
The robot is a modified walker (the kind that people use for
mobility). It's got a touch screen, computer, and laser
rangefinder bolted on, as well as motors/encoders for the rear
wheels. Its main activities during its lifetime were helping
people find their way around an assisted living center (the one
in Plum) and parking itself out of the way and returning when
called. (For the latter, apparently crowding of walkers and canes
near chairs is a major cause of falls.) Other people who worked
on the project include Sebastian Thrun and Judy Matthews.
Another one at NREC?
From: Lee Weiss
Having been a founding member of RI and being here since the
start in 79', and having designed and built the (I think) first
robotics lab/system in RI - and more since then (e.g. first
robotic additive manufacturing facility, anywhere) - I have
numerous 'early day' stories and pictures. Also, while these
early systems were built around (mostly) commercially available
bots but with custom grippers (the first one of which used in the
'flexible assembly cell' is still sitting on a shelf in my
office), I built a 'robot' (or, more appropriately, programmable
machine) that was the first bioprinter designed to print in
situ, directly into the body.
From: Red Whittaker
Kevin Dowling is the key to the authoritative story that you seek. He
participated broadly and cross-pollinated people, labs, teams, projects
and technologies. He knew (and retains the memory to still know)
everything. Kevin is encyclopedic. I've cc'd Kevin for eyes on this
topic.
The Microbot Mini Movers were the first commmercial educational/research
robot at CMU. These were purchased and broadly utilized in research,
teaching and demonstrations.
David Bourne had industrial manipulators. Other manipulators were used
for visual servoing research. American Robot Merlins came along in the
early 80's. They were produced in Pittsburgh by Romesh Wadhwani (cmu
alum). The big thing in the day was to hack the
industrial robot to develop better controllers, sensors,.. and interface
directly from computers.
The first cmu-born mobile robots that exhibited real technology were
Pluto from Hans Moravec's lab, the rotating-sonar robot (not the actual
name) from Jim Crowley's lab. There are great videos and stills of
both. Crowley's is really photogenic due to the motorized sonar that
continuously rotates like a lighthouse. Way cool. Still one of my
all-time favorites. I protected all these for many years, but Crowley's
was taken, canibalized and destroyed. The last Pluto existed until 9
years ago, when it was dumpstered during a purge. The robots and
technologies were painstakenly developed over time. It would be
straightforward to replicate them given our deep knowledge and that the
key people are still around. These are iconic, worthy, important enough
and easy enough to replicate if it mattered. A concurrent development,
and the first to leave campus into the world was our Remote
Reconnaissance Robot for Three Mile Island. There were a pair. One
cleaned up the basement of the TMI accident. The other is in
Pennsylvania's State Museum in Harrisburg. Maybe they'd give that one
back.
From: James Montgomery
We have the original machines from the UGCV project - Crusher, Spinner,
and Spinner's original prototype, Virgil. There are also (three?)
PerceptOR vehicles and a Black Knight tank.
I just remembered that we also have a couple of the Gladiator prototypes
here in our North Bay.
From: Gregg
I may be able to locate some of the ones I built.
Pluto (designed by Karen ???, and modified by me) moved to the Boston
Computer Museum many years ago.
Neptune (and two darker blue copies made by other labs) were in the
catacombs under the highbay as recently as a year ago.
Uranus was in the lobby if SeeGrid where Hans Moravec is about two or
three years ago.
Test Inspection Platform (TIP) for table top aircraft skin inspection I
have most of in my basement.
Crown Inspection Mobile Platform (CIMP) for fuselage remote visual
inspection had parts stored on the upper platform in the sub basement
b-512? That John Dolan oversees. Some parts had been in the Electric
Garage, and Ben Brown may know.
Three Robot Sensor Boats are probably still in B-512 (John Dolan again)
but may still be in use.
YODR (used to carry scissors onstage for the Gates ribbon cutting) may
be in John Dolan's Floor 1 or A lab. Alan Guisewite may know.
BarbieBot and Fullabot are in Howie Choset's lab. May still be in use.
I made a robotic conference room chair for Intel which is probably gone
but Illah may know about it.
Two robot page turning arms were at Google. Illah again on these.
An ever-expanding fleet of PF3-30 robots are in use by Howie.
From: Kevin Dowling
David is likely right about the first robots designed and built
at CMU through Red was building systems for TMI about that time.
Here's my recollection of the robots.
The first robot in the Robotics Institute was a Seiko Model 700 -
a pneumatic-powered bang bang robot used in the Flexible Assembly
Station. I pulled a trick on Raj with this by setting is up to
throw a large bolt. I had mounted the first vision system on the
Seiko and then set it up so when Raj came by, I said 'Check this
out, Raj' and put a large can on the floor nearby. The robot then
picked up a bolt, swung, and tossed the bolt into the 'randomly'
placed can. Raj's eyes bugged out for a second....
PUMAs - we had three PUMAs in the Flexible Assembly Station that
was the focus of the work by Art Sanderson and his graduate
student (at the time) Lee Weiss. One of the PUMAs was also used
for real-time tracking using Kalman filters and a vision system.
It worked but it was painfully slow.
David's robot was labeled Copperweld and there's an interesting
story there. David was trying to control it by voice and needed
to interface to the robot and get inside. Somehow Copperweld must
have heard we wanted to modify it and they sent someone in to
replace the main processor board. They wanted to replace with a
board where the chip numbers were erased. He arrived late morning
and I had someone take him to lunch. I opened his briefcase and
copied down all of the labels for the chips and then put
everything back. He never knew. and we were able to build the
interface.
Jim Crowley had the heavily modified Hero robot. It used an
upward pointing Polaroid sonar sensor and a right-angled cone
that spun, giving a range map of the surroundings.
Hans Moravec had Pluto, a very complex and clever 6-DOF wheeled
system. I worked on a stereo slider for that as an undergrad.
Pluto never really worked well - overconstrained systems tend to
be difficult.
So Neptune, a simple three-wheeled system, became the workhorse
for Chuck, Larry Mathies, Alberto Elfes, Rich Wallace and others
using stereo camera and ultrasonic ring setups. It navigated the
room after many trials. Best story was that during a series of
runs - might have been Chuck or Larry's code at the time - it was
consistently off by about 10%. The error was finally determined
to be a conversion issue - sort of. Ten inches to the foot; a
metric foot! Greg Podnar then design Uranus, which used Mecanum
Wheels to provide 3DOF in the plane. Very cool robot.
Red was building TMI 1 and TMI 2 for Three Mile Island.
Teleoperated devices made with polished stainless throughout to
facilitate washdown. Counterwind tethers - very cleverly done.
Then Workhorse, a massive teleoperated system that was never used
at TMI and is in the NREC extrance area. It could do a lot of
things.
In between TMI machines was the Terregator. Six-wheeled
terrestrial navigator. Was the plucky robot that went on
sidewalks, inside mines, and was used for years. Saw several
upgrades and renovations. I breathed a lot of T'gator's exhaust
while folks ran software on the mainframes elsewhere. Lots of
stories around Terregator.
Takeo and Asada's DD arms were impressive. Unfortunately Harihiko
Asada mis read the specs on the main joint for DD1 and it always
had trouble lifting, The new blue DD2 was a beautiful design by
Don Schmidt and, I believe was Pradeep's thesis.
Navlab 1 in 1986 was the first of the Navlab family. Starting
with a Chevy truck chassis, I swapped out the transmission, added
a large shell and packed it full of racks, generators, computers,
and sensors including the ERIM laser scanner. An amazing vehicle
that was ponderous but was a 1 robot army for theses, papers, and
the star of more than a few articles and TV programs. It was
followed by numbers 2-6 ranging from HMMWVs to sedans and mini
vans.
Ambler was another large and ambitious program incorporating a
several meter high novel legged walking machine and a whole new
program of planning and navigation. One of my favorite robots I
ever worked on.
Tessellator was a NASA project for�a omni-wheeled�system to
service the underside of the space shuttle. Big robot - fun trip
to Kennedy Space Center which is the subject of another story.
There were excavators, manipulators, my early snake robot and
many others that will come to me. I have pictures and MJ still
has a set of slides of many of these.
Lynch's one-joint dynamic manipulator (1997, Lynch, Mason)
MLab robots:
The Palm Pilot Robot Kit.
The Mobipulator.
The MLab Simple Hand.
Ballbot
Illah had a museum exhibit where visitors got to play the part of
earth-bound robot drivers, while little MER emulators explored emulated
Mars landscape. Very popular.
Illah also had the insect telepresence robot.
Also one of the first robot guide robots.
Reids internet bot. Xavier.
Commercial robots at CMU
* The Microbot Mini Movers were the first commmercial
educational/research
robot at CMU. These were purchased and broadly utilized in research,
teaching and demonstrations.
* Flexible Assembly Station: The first robot in the Robotics Institute
was
a Seiko Model 700 - a pneumatic-powered bang bang robot used in the
Flexible Assembly Station. Kevin Dowling
Lee Weiss and Art Sanderson built a flexible assembly cell to
install IC's into boards. That was based around 3 Pumas and special
hardware.
I was using a bang-bang robot that I controlled with voice.
David Bourne
* Copperweld: David's robot was labeled Copperweld
* Commodore 64/Hero robot. Jim Crowley had the heavily modified Hero
robot. It used an upward pointing Polaroid sonar sensor and a
right-angled cone that spun, giving a range map of the surroundings.
Crowley's was taken, canibalized and destroyed (Red).
Jim Crowley built the first mobile robot based on the Commodore 64/Hero
robot. But he automated mapping of the workspace.
* American Robot Merlins came along in the early 80's. They were
produced in Pittsburgh by Romesh Wadhwani (cmu alum)
* Cincinnati Milacron T3
From: David Bourne
Paul Wright and company were using the Cincinnati Milacron robot
in Mechanical Engineering to do all sorts of different experiments.
Can you use a robot to fill a gas tank? Paul and Mark Cutkosky did
some nice experiments with automated grinding.
Note to self:
The RI published an annual research review starting from year 1.
The CMU Archive has online copies.
