Sensory Overload did not fair so well at the 2006 AHRC Robot Rally Mini Sumo contest. With a record of 2-2
the robot ended up in 3rd or 4th out of 8 or so competitors.
Updated: 6/10/06 Sensor Relocation : After many trials it was determined that having the sensors mounted on the top of the
robot was not going to work. With a little force an cutting the sensors were moved inside the chassis
and a hole was cut into the front wedge to allow an unobstructed view of the opponent.
Old Sensor Location
Here is a shot of the original location of the analog IR obstacle sensors. There were a lot of reflections
from the top of the robot that made the sensors unreliable at best.
New Sensor Location
As you can see the sensors are now much closer to the ground giving a much better performance as compared
to the original position.
Sensory Overload was designed and built to compete in the 2005 Botlanta Mini Sumo contest. Unfortunately
the contest was canceled :-(.
Sensory Overload was my first attempt to build a mini sumo robot. I had built other robots but had never built
a robot with the intention of competing with it. I decided to take a more exotic approach. Mini sumo rules
dictate that the robot can be no larger the 10cm by 10cm but there is no restriction on height. So I decided
that I would do what I could to exploit this loop hole. The robot is placed into the ring on its back and
flips forward once the round is started. This method allowed the robot to be longer then the 10cm restriction.
The robots chassis is made entirely of Plexiglas. Plexiglas is not the easiest material to work with but I really
like to see the mechanical inner working of my robots. The side panels are 1/4" and the rest is 1/8". The sides
were drilled and tapped to accept 4-40 screws that hold to robot together.
The robot uses two Solarbotics GM3 motors for the drive. The motors are not that powerful but on a less then
500 gram robot they get the job done. They are also really cheap at less then 15$ a pair. The motors are
screwed together with two pieces of 4-40 threaded rod. There is a 1/4" piece of Plexiglas in between the
motors that mounts to the bottom of the chassis. The motors second shaft had to be carefully removed in order
for the motors to fit properly in the chassis.
I decided not to use the standard thin wheels that normally mount to the Solarbotics motors as I wanted to get
more traction. I created a custom adapter that matted a Tamiya Racing Tires to the GM3 output shaft. The parts
were created in the free
software and were laser cut by
In order to see ones opponent sensors must be employed. The robot uses two Sharp GP2Y0A21YK analog IR sensors.
The sensors output a voltage according to how far away the object reflecting the light back is.
The robot uses two QRE1114 IR emitter/IR Photo transistors to detect the line. The sensors have a fairly
limited range so they have to be placed very close to the ground. There are a total of four line sensors on
the robot but the back two didn't offer anything as useful as by the time the sensors triggered it was too late
anyway. The intention of the rear line sensors was to allow the robot to take evasive action if being pushed but
the code for that never got created. The grey wires with the masking tape on the end is the ICSP port for the
micro to allow easy in system programming.
To make the robot more exciting to watch I built in an LED board that has four blue LEDs and eight RGB LEDs.
The blue leds indicate what direction the opponent in sensed and the RGB LEDs act as sort of a mood indication.
When the robot is on the attack the entire robot turns red and when the robot is lost it turns blue. Green
is used to indicate that the other robot is found but has not been zeroed in on yet.
The robot is driven by a PIC16F876 @ 20MHz. The micro monitors all the sensors and takes the appropriate action
when a sensor is triggered. The algorithm used was a "If you see something, run into it" type brute force
thing. The analog readings from the sensors are averaged to determine if the where the opponent is and how
far away. A PCB was created for the circuit cards as most of the parts used in the robot are surface
mount. I was looking for a reason to use surface mount parts as I had not used them before. Wish I would have
gotten some pictures of the board before it went into the chassis. It was my first double sided board and it
turned out pretty well.
Sensory Overload vs. Kleenex Box Round 1
Sensory Overload vs. Klenex Box Round 2
Sensory Overload vs. Kleenex Box Round 3
Sensory Overload Data Sheet
3.9 in(W) x 5.5 in(L) x 3.875 in(H)
1/4" and 1/8" Plexiglas
C written in the SouceBoost IDE
7.4V from a two cell 2200mAH lithium-ion battery pack.
A linear regulator steps the voltage down to 5V for the PIC.