■ FIGURE 2. A labyrinth toy.
own functions (tracking the sun for optimal
energy collection), but also for propulsion and
steering of an unmanned vehicle. Fortunately,
we didn’t have to design the mobile unit. We
only had to design the solar tracker, with the
provision that it have enough energy left over
(stored in a battery) so that it could power a mobile unit.
■ FIGURE 1. Solar Tracker on my robotics shop (i.e., basement)
It was the intent of this course to give us a taste of
real-world engineering; as such, all decisions were based
upon things such as safety, cost, manufacturability, power
consumption, component strengths/weaknesses, and
impact on the environment. In order to quantify these
multiple factors, engineers use a tool called a decision
matrix — nothing more than a chart which uses weighted
values to compare the merits of different candidate
solutions to a problem. All of the weighted values are
tallied, and the candidate with the highest total wins.
For example, we had to select an actuator to position
the solar panel so that maximum surface area faced the
sun. What kind of actuators should we use: stepper
motors, servos, or regular DC motors? I was biased toward
steppers because I’m a big nerd and had already spent a
lot of my own spare time (more than I care to admit)
designing a stepper motor driver and really wanted to
implement it. But were stepper motors really the best
solution for this project? Having never used servos before,
we researched them extensively.
Many tests were run on all three candidates. Our solar
panel only produced 20 watts of energy, so actuator
power consumption was heavily weighted. Torque and
accurate positioning were also paramount. As it turned
out — much to my dismay — servos were determined to
be the best candidate solution. They are internally geared
down, providing adequate torque to move the solar panel.
Servos also have internal circuitry which provides negative
feedback for accurate positioning — saving us from having
to design an external feedback device (such as an optical
encoder). They also consume less power than stepper
motors. So, though I didn’t get to use my custom stepper
motor driver, I did learn a lot about servos.
Decision matrixes were also used to determine the
best mechanism for positioning the solar panel. We
considered moving a big lens to focus the sun’s rays on
the panel. This method would have focused a lot of
sunlight (solar energy) on a small fraction of the solar
panel’s surface, but would the increased local power
density be worth the unused real estate? We also
considered pivoting the panel on the end of a rod, which
came to be termed the “joystick” method. We finally
settled on using a frame-in-a-frame method, like the old
labyrinth toy (Figure 2). This provided two axes of motion
and the simplest mechanical design.
After months of decision matrixes, Gantt charts, and
lots of other boring stuff, we were finally ready for the fun
part — prototyping!
West Virginia University Institute of Technology
THE “BUG EYE”
If you are interested in studying engineering, I personally
recommend WVU Tech! In 2007, WVU Tech was named as
one of the nation’s top 100 Undergraduate Engineering
Programs by US News and World Report. With an enrollment
of only 1,452, the faculty/student ratio is very low — which
means a lot of individual attention.
Okay, remember the three basic elements of a good,
interactive machine that I mentioned earlier: input,
processing, and output? Think about it like this. Say you
want to pick up an apple. First, you look around (input).
Light is reflected off of everything around you and into
August 2009 23