CS 490 Special
Projects in Computer Science: REAL-TIME ROBOTICS
Winter 2002, Instructor:
Jeffrey Horn
- Digital Braitenberg Vehicles
- Members:
Carey Stortz, Chaplin Cinelli, Bob Lindquist
- Basic idea: Design a whole
new type of very simple, but evolvable, robots. Using just a few
logics gates, AND, OR, and NOT perhaps, we can wire up ever more complex
stimulus-response behaviors. I thought we had invented this (the
digital version of Braitenberg's original analog vehicles) here at NMU,
but then I found a paper from MIT describing the same idea (see the
paper in the cabinet in 1207). But they never really built them,
and plus I want to use them to teach computer architecture, AI,
artificial evolution of robots, etc. I'll explain more when I get
back. For now:
- Look over the paper from MIT.
- Take out the brand new electronics kit from Radio Shack (in cabinet)
and build a few test circuits from the book.
- Then try building a "Vehicle behavior" circuit, such as
light avoidance, using just the NAND gate chips (or NOR or whatever) in
the kit. Use LEDs to simulated turning on the motors, and buttons
as input to simulate the sensors. I just want you to check out the
kit and to get a feel for designing and building logic circuits with it.
- LAPTOP BOB
- Members: Michael
Sanford, Branden Hoiska, Jeremiah Last, Mike Oman
- Basic idea: is to build a
robot around a ThinkPad, using tele-presence for control, in order to
show off the capabilities of an NMU laptop and to have a nice demo bot
that is always working and always avail. for all of those
"events" we are constantly being asked to bring 'bots to
(e.g., the academic fair, open houses, etc.). 'Bot would consist
of a mobile ThinkPad with tele-presence. The driver's face would
appear on the laptop screen via wireless ethernet, he/she would drive it
via a joystick or mouse, and a video camera on the laptop would transmit
(perhaps via the 802.11 connection) back to the driver's PC.
Two-way audio included. So the idea is that anyone, via the
internet, could drive the laptop around a cocktail party, say, and feel
like they were "there", starting up conversations with people,
etc.
- Go get the old Max Base 'bot from the robot lab. Check it out as
a possible base for our roving laptop 'bot. Could we mount a
thinkpad on here easily?
- Get the Orinico wireless card from Keith and install it in a laptop
(I'll work on finding an extra one if we decide to go ahead with this
approach). It should work in the atrium and maybe
1207).
- Try out video feed from the 'bot to the internet. Grab the USB
video converter I left for you in the cabinet. Unfortunately, I
can't find the disk with the driver. So if you have time, got to http://www.x10.com
and look around for the driver and instructions on how to serve up the
video feed on the web. Or, do a google search on "USB
video" and look for some decent products (should be around
$50-100). Maybe we should get a better one. Can you think of
any faster way to get video into the laptop and out through wireless
Ethernet? USB limits us to 11 Mbps. Use the VCR output for
testing the video input, if you can't find the cameras.
- High-speed Tele-presence
Rover (the "car")
- Members: Ben
Maki, Joe Schmeltzer, Kurt Payne, Nate Rabe
- Basic idea: Perfect the high
speed rover from last year! This has the potential to be a very
cool, high-visability project, with much funding potential (ask me about
the volcano sampler!). But we need to smooth this system out (no
more vomiting!). So let's see shat we can do.
- First, get out the poster session on this project and read it (click here,
or read the hardcopy in the 1207 lab, stuck between the cabinet and the
bookcase).
- Then look at the rover in the closet, and look at the truck
base. Notice how overloaded the car has become. Would the
truck base be better? Would it be strong enough? Remember,
we might want to go outside and off-road with this eventually!
- If you don't think either is enough, grab the Tower Hobbies catalog
out of the cabinet, and find us a decent kit: four wheel drive,
maybe a metal chassis! You may consider nitro (gas) vehicles (we
can always substitute electric motors; I do NOT plan to drive nitros
inside NSF!). You can also visit www.towerhobbies.com,
although I have always found their web site lacking.
- The "Northern Segway"
(a.k.a., "yooper scooter")
- Members: Keith
Rutkowski, Scott Gowell, Lyle Hanson
- Basic idea: Take an electric
scooter, and turn it into a self-balancing 'bot. This might be a
great application of the real-time Linux running on the Megatel, a great
demo of the real-time capabilities, if we can get the thing to stay
up! It would look pretty cool too!
- Get the scooter out of the closet and take it for a spin.
BEWARE! DO NOT OVERDO IT! If you get us into trouble by
running people down in the hallways, well... So ride it discreetly
and when the hallways are pretty empty. Or take it to the
tunnel. Don't ride it outside (it will get dirty and besides, it
doesn't have much power to handle slopes or rough surfaces, too wimpy).
- Take it apart. Get screw drivers from small robot lab (look in
table drawers).
- See if you can remove (sniff) the handle bars and steering column down
as low as possible (do not destroy anything! You never know when
we might want to put the steering column back on; there should be a
non-destructive way to do it, in other words
"reversible". I hope. Try it.)
- See if there is any way to bypass the necessity of a running start (if
you haven't found out yet, the motor will not engage unless you are
already moving fairly quickly; but is this because of a limitation of
the motor drive circuit, or is it a "safety feature"? If
for safety only, REMOVE IT!). I don't want to have to push start a
'bot!
- We will need an "inclinometer" sensor to tell us when she is
tipping. The only one I could find is on the sheet in the
cabinet. Look at it. Got any better ideas? Do a web
search. See if you can find anything else cheap, simple, and with
a "digital interface" that we can hook directly to a
microprocessor. Send me the links of ANY candidates.