Twin Tweaks ...
THE SCRIBBLER'S IR SENSORS ARE READY TO
STARE DOWN SOME OBSTACLES.
compared the Scribbler to the Mark III. Surely other
tinkerers have had similar burning questions about the
artistic bot, so the answers could likely be found on a
robotics forum somewhere. However, we preferred to live
by the sage mantra of Reading Rainbow’s LeVar Burton,
so we wanted to see for ourselves instead of taking their
word for it.
We pulled out our trusty multimeter to do a simple
test that would determine if the sensors were analog or
digital. We pressed the contacts against the ground and
signal pins of the infrared receiver and moved an obstacle
in front of the bot. If the multimeter gave a reading that
jumped from zero volts to five volts once an obstacle got
close enough, then it was treated like a digital input.
If the multimeter gave a reading that gradually increased
from zero to five volts as an obstacle approached, then
the sensor was being used as an analog input. The
results of our test were somewhat difficult to interpret
given the difficulty in placing the contacts only on the
individual legs of the infrared receiver, but we did only see
dramatic jumps from zero to about five volts and no
To seal the deal, we inspected the Scribbler program
itself. The specialized Scribbler GUI — even with all of its
colorful building blocks and images of brick walls that
flagged the blocks dealing with obstacle avoidance — was
about as forthcoming with technical details as Raj
Koothrappali is with conversation around attractive women.
Thankfully, the Scribbler program offers the option to see
the code in Basic form, and that is where we found our
definitive answer. The IR sensors were treated as either true
or false — the hallmark of a digital input. Now that we had
solved the mystery, the next question was what were we
going to do with that prized knowledge.
A Sharper Image A LOOK INSIDE REVEALS LITTLE ABOUT THE ENIGMATIC SENSORS.
INVESTIGATING THE NATURE OF THE SCRIBBLER SENSORS.
68 SERVO 05.2012
One of the challenges in assessing the rangefinder
sensors was to devise a test that would allow us to
accurately track the results. Very often, rangefinding
sensors are used for obstacle avoidance, usually in the
context of a maze challenge. So when the sensor works,
the robot doesn’t hit the obstacle. Sometimes when the
sensor doesn’t work, the robot might miss the obstacle
anyway out of pure luck or happenstance. So, what we
needed was an obstacle directly in the path of the bot
and a behavior that would unambiguously indicate
sensing of the obstacle. For the Scribbler, this was an
easy task. The LEDs could light up when an obstacle was
sensed, and we could have the robot back up. A similar
behavior could be programmed into the Mark III, minus
the helpful LEDs.
The other challenge with devising a test was to
determine a way to fairly compare the digital inputs on
the Scribbler with the analog rangefinders on the Mark III.
The Mark III uses two 2YOA21 Sharp rangefinders. The
2YOA21 sensors are the de facto successors to the