30 SERVO 11.2017
Many vessels are equipped with emergency beacons.
Aircraft and seafaring vessels both have position beacons
and even an infamous “black box” that records state data. I
wanted a simple beacon that would help me find my
downed quad in the simplest way possible.
First, I considered a GPS beacon that would send a text
message with the final resting coordinates via the cell
network. That seems over-complicated and assumes that
there will be cell coverage — not always a given in remote
areas. It also is a complex enough system that several
things could go wrong, including the shearing of an
antenna during a crash.
Next, I considered a Wi-Fi solution that would just
require me to be near the vehicle. There are still concerns
about the antenna surviving the crash, and I would need to
be pretty close to the craft in the first place. It also
consumes a lot of power, meaning that it wouldn’t be a
beacon for very long on a small battery.
Large batteries add weight and I want to remain
independent of the quad’s power system in case that is
what caused the crash. With that solution out as well, I
decided that something following the KISS (Keep It Simple
Stupid) principle would be the best.
The solution I settled on was indeed dead simple: a
piezo buzzer. Once the quad detected that it was in free
fall, I wanted a loud tone set to be broadcast every few
seconds. Sound carries surprisingly well and our ears are
great direction finders.
I couldn’t quite leave the system that simple, so I
thought it would be nice to have a quick and easy arming
mechanism, a visual status indicator, and a low battery
warning. With these features in mind, I needed to figure
out how to detect free fall and then get to work.
Detecting a Crash
The first order of business was determining when the
beacon needed to be activated. I didn’t want it to
accidentally turn on during normal flight, requiring a trip
back to home base to turn off the beeping. Failure to
activate during a crash would be just as bad.
A few sensors come to mind. I could sense the altitude
and activate when the rate of decent exceeded some
threshold (also known as the derivative for those versed in
differential calculus). This would most likely work, but
requires measuring the surface pressure, filtering noise on
the signal, and a few other annoyances. I could use a gyro
and trigger on high angular velocities, but not all crashes
have significant spin components to them.
I finally settled on the accelerometer. The one thing
most crashes will have in common is a period of free fall. In
fact, I think it would be pretty hard to have a powered
downward acceleration (flying with the quad upside down)
that is stable for any
length of time. Even if
you hit an obstacle,
descent is in free fall.
Now that we
accelerometer is what
we will use, we need
to figure out how to
interpret the data it is
giving us and
determine when the
craft is falling out of
the sky. To do that,
we need to first
what acceleration is
and what the
accelerometer tells us
acceleration is the
Build a Drone Crash Beacon
I’ve been toying with ideas about using drones to help with agricultural evaluation, but keep having the
same vision of my quad plunging down into a corn field. After hours of searching, I doubt that I would
be able to find it (there are such things are corn mazes, remember). In fact, it would probably sit out in
the field until harvest and then be mercilessly ground up by harvesting equipment. Who knows what the
aging and weathered battery would do. Not a situation I want to be in. So, I’ve made a beacon that will
make finding a downed copter much easier.
By John Leeman
Figure 1: Accelerometers will report
the accelerations they experience
in the X, Y, and Z directions.
Knowing that gravity always
accelerates things towards the
we can resolve the relative
orientation of the sensor.