mlml logo

nsf logo Powered by Blogger.
Tuesday, November 20, 2007: Electromechanic Wonderland Miscellany
This week will be a lot of bench time for me as I debug some thruster power issues, recalibrate sensors, clean up our controls software, and add some inertial measurement features. So while everybody else is out breaking their backs in beautiful high twenties weather I'm getting fat eating cookies, electrocuting myself, and putting the Mr. Coffee through a stress test.


(continued...)


Being a physics dork, I'm pretty excited about our inertial sensors. These micro-electro mechanical systems (MEMS) devices are tiny enough to fit on a standard microchip board and give us feedback on how SCINI is oriented ("Which way is down?") and how fast we're turning ("Wait, are we moving?"). Our acoustic navigation system gives us positions in space over time, so we can figure out exactly where we are and sometimes roughly how fast we're moving, but it's accuracy and speed (updates every couple seconds) make it difficult to tell what's going on right now, and gives us no information at all about SCINI's orientation with respect to the ground. In good conditions we can look around for a dive rope or even back at the hole to figure out where we are, but sometimes the ground is sloped or we are flying in pitch black and have no reference.

The accelerometers measure the force being applied on SCINI; sometimes this is due to our thrusting about or handling on the surface, but mostly it's due to gravity (or more precisely the buoyant forces keeping us from free falling at the acceleration of gravity). By comparing our acceleration along three different axes (forwards/backwards, up/down, left/right), we can calculate the gravitational force vector in SCINI's frame of reference (aka, which way is straight down relative to which way the ROV is looking); because on human scales this direction is pretty constant, we then invert this vector to give the pilot SCINI's orientation vector in their frame of reference (aka, which way the ROV is pointing relative to the ground). Because we don't have a compass (the magnetic fields are crazy around here; magnetic south is actually north of us) and gravity is symmetrical around a vertical axis (you can't "feel" which way is north or west like you can up and down), we don't know which direction we are pointed on the ground plane (for navigation), but we know our pitch and roll tilt, which are more important for maneuvering.

The tiny circuit board sticking out perpendicular from the main microcontroller board contains all of our inertial sensors

The accelerometers are pretty simple conceptually: there is a tiny lever with a weight on the end which gets flexed when a force is applied on it; the sensor measures how far the level is flexed to determine the force. Our gyroscopes, which measure the rate of rotation, are a little trickier. A classic gyroscope is a spinning weight suspended so that it can twist freely; the weight wants to keep spinning the same way (think of the bicycle tire trick), so you measure the orientation of the weight to determine how far you have rotated relative to it. It's pretty tricky to suspend a large enough weight and keep it spun up in a cheap way, however, so instead we use sensor based around a tiny vibrating mass: when twisted the coriolis effect perturb the path of the mass to one side or the other, and this is measured to determine the speed of rotation (aka angular velocity). If you're interested in all that, check out these HOT HOT links: white paper, wikipedia, manufacturer description.

And now... filler photos! Leave a comment if you are curious about any of these; I realized we had anonymous comments disabled, the process should be easier now.

Icefish skull

Quiet Crary lab hallway

Nudibranch hanging out in the Crary aquarium

In case nobody else has mentioned it, Antarctica is a harsh continent, and also doesn't care. These truths tragically hold true for invertebrates as well as us boney types, even wise beasts like this octopus that got caught in a fish trap over the weekend. The effects of decompression coming up from more than 140 meters (400+ feet) were too much for the poor fellah and he didn't make it through the night, but for a while he ruled the fish tank benevolently, putting feisty bernacchii in their place with a smack from his mighty tentacle and cycling through skin colors faster than a cabernet singer. He will be missed.





Comments:
Post a Comment!

Subscribe to Post Comments [Atom]



<< Home
Feeds
Subscribe to Posts [Atom]

Archives
This material is based on work supported by the National Science Foundation under Grant No. ANT-0619622 (http://www.nsf.gov). Any opinions, findings and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the National Science Foundation.
Webmaster