I got a video up showing the simple Segbot 555 timer based balancing
robot. If you're going to get into balancing robots or working with
stabilizing complex dynamic systems such as flight control, this
robot is a good place to start.
The 9 volt battery you see is just to add weight towards the top.
The servo used was a Cirrus CS-71 modified for continuous rotation.
This is about as close as it gets to simplicity for the mechanics
and electronics in a balancing robot.
Don't be intimidated by the notion of building a balancing robot! If
you can build BEAM heads then you can build this robot. The
schematics are in the file section.
I also uploaded some schematics showing some work I've done to see
how much processing power can be pulled out of a single transistor
robotic system. This is linear motor control and I present it as a
novelty only (please, don't flame me, it's a novelty). It does,
however, work quite well with smaller motors and has very dynamic
and chaotic behavior due to the high system information capability
I was able to get 4 distinct subsumption behaviors out of a single
transistor control system: scan for light, follow the light, avoid
side obstacles and stop for front obstacles. This means that a
single transistor control system can show emergent behavior!
The amount of light intensity will determine how hard that the rover
turns. The PID option gives information on light direction, light
intensity and rate of change (due to the derivative function).
You can get rather complex patterns if you let the robot orbit
around a light source. Lay down a large white board, set up the
light, have the rover drag a dry erase pen and you can dra out
chaotic orbits. One set up gave a nice chaotic rosette.
As a light orbiter the single transistor system fulfills the three
basic requirements for a chaotic system: sensitive to intitial
conditions (you can get different patterns depending on the starting
point), increased entropy with time (the patterns become more and
more complex) and a posistive Lyanpunov exponent (it's a divergent
system but a close orbit).
As a orbiter the 1T rover also demostrates a basic chaotic
associative memory function since it tends to settle down to a
unstable periodic orbit even with system perbutations.
The mosfet version isn't as interesting due to a very small linear