If you or your business would be interesting in sponsoring the
FIRST robotics teams involved with this project and making future
projects possible, please contact us.
Disclaimer: DIY Segways, like any large machines, can be hazardous if appropriate safety precautions are not observed. DIY Segways are particularly dangerous because they often lack the redundant safety features of commercial Segways. This
site is intended for informational, not instructional, purposes.
Attempt/build at
your own risk. Be careful!
Summer 2008 is here, and you'd be correct to suspect that a new project is also in the works. Check out the Cap Kart, a unique electric go-kart with some special features...
Also, check out these awesome new self-balancing projects from around the web:
The DIY Segway made a trip out to Wayland HS, where it was put to the test by some of the designers, who also happen to be the more experienced riders. The video features some of the things our Swapfest appearance did not get to show off, including performance on an incline, performance on grass (not great, but with 12.5" wheels, we weren't expecting to do much off-roading), turning in place, and clearing a 2.5" tall camera, with 2.5" to spare. Can your Segway do that?
We are a group of students from Wayland High School, John D. O'Bryant
School of Math and Science, and
Cambridge Rindge and Latin School working with four MIT students. We all have worked on FIRST Robotics teams and have some
background in engineering. During the FIRST Robotics competition in
Atlanta this year, we had an idea...
FIRST Robotics is a lot of fun. However, it occupies around eight
weeks of any given year, usually. Eight weeks out of fifty-two leaves
plenty of time for other projects. In Atlanta this year we were
brainstorming what strange contraptions we could throw together over
the summer. Although Ed, a technical instructor at MIT who works with
our team, suggested a fountain, Shane had been saying it would be easy
to build a Segway and we called
him on it. Besides, you can't ride a fountain...
When looking at a scooter on two wheels automagically balancing
itself, it may seem an almost impossible feat. However, the way a
balancing scooter works is rather simple.
Think of it this way: When you stumble, you don't (usually) let
yourself fall flat on your face. Instead, you put a foot in front of
you to "catch up" with you and get below your center of gravity. The
segway works in the same way -- when it detects you leaning forward,
it moves forwards to catch up. When it detects you leaning backward,
it moves backwards to catch up.
When designing a segway, there's a lot to consider. You have to
make sure that the motors you're using have enough torque to move the
weight of a person around and keep them from falling. You have to be
careful that the motor shafts won't snap when a person hops onto the
scooter. And of course, you have to make sure it can go fast and look
cool. In addition to all that, we aimed for a few other features in
our early design process. We wanted to:
Have it weigh less than 50 pounds with the battery
Spend less than $1000
Be able to ride our scooter through doors
Have "lean steering"
A "bonus feature":
For more design-related stuff, check out the notes from our design meeting.
Once we designed our DIY Segway, we ordered parts and cut
our aluminum baseplate on the waterjet in MIT's Hobby Shop. It took us three to four meetings to fully put together our segway,
although what required the most tweaking was the damping on our lean steering mechanism. We tried a few different
methods before we finally got that right. As of now we're using pieces
of lexan that act as leaf springs.
The DIY Segway is controlled by a PIC
microcontroller board based on the Machine
Science XBoard. It interfaces with a computer for programming
the chip or viewing sensor values in real-time. Note that although
a laptop is used for debugging, the segway itself is controlled
entirely by the microcontroller. Its "deadman's switch" is a
magnetic card reader, as pictured below. For sensors, we used an
ADXL203EB
accelerometer to sense acceleration due to gravity -- which is a good
way to measure the angle -- coupled with an ADXRS401
gyroscope to measure the rate of angular rotation (how fast the
person is falling forward). The Apple
iPhone also uses an accelerometer to determine when you tilt it...
but you can't ride an iPhone!
All of the code on-board the DIY Segway was written in C and compiled /
loaded with Machine Science's
IDE. The heavily commented source code can be viewed nicely in a
browser here and downloaded here. If you have any questions about the
code or suggestions for improvement, feel free to direct them to this e-mail address. The
code's main loop currently runs at about one hundred times per second
(100 Hz), which is more than adequate for keeping a person balanced.
It also uses the onboard radio to transmit variables to a laptop for
real-time debugging, as shown in the screenshot below.
On August 19th, Shane and myself (Cam) went to MIT's Swapfest and
encountered somebody riding a commercial "Segway Transporter". Since
we built the segway less than a five minute walk from where Swapfest
is, we decided to run back and bring our DIY Segway outside for its first
public outing. It went very well. If you're impatient or bored, you
can skip to the video below, but if you want to read on then here's
how our DIY Segway stood up to the commercial Segway Transporter:
DIY Segway
Segway Transporter
Top Speed
Theoretically 11mph
12.5mph in newer versions
Steering Mechanism
Lean steering
Twist handlebar in older segways, lean steering in newer
segways