Question: Question: Would the Thermal Fishing Bob concept work as a SeaGlider robot?

I believe it would be better to just store the data and use dead reckoning between fixes. That assumes you get GPS on each surfacing.

In that case, perhaps a better question would be about using the sensor pod as a platform for RIFFLE or other Open Water tools? The only problem with the GPS method, is it would either have to be highly accurate, or the sample area would have to be wider. This would be a problem because while the Seaglide can move autonomously, it can't NAVIGATE autonomously... A wide sample area would either require a really long leash or the addition of some kind of navigation capability. If it already has a GPS, then NAV would not be impossible and could probably be ported from Arduino-UAV sources like MultiWii. But AFAIK, it really can't control its direction very well in its present configuration. I don't think adding such a capability would be impossible either, but would likely require quite a bit more testing...

I reviewed the design and there is no port and starboard control. That could be accomplished by changing the center of gravity to the left or right. Nor does it have GPS. I don't believe high accuracy is needed though. By the way the GPS will only work on or near the surface.

If flow/water speed and pressure/descent was being measured too then those measurements could be used to generate approximate locations between GPS waypoints.

Probably just pressure would be sufficient. I am not certain but there is probably a constant relationship between pressure change rate and speed/distance. That will probably require individual calibrations which may be accomplished via GPS fixes though.

I just remembered the other sensor that was being used was a digital compass to get both heading and pitch.

i wonder if it really would make sense then to base the navigation controls off of a flight controller IMU like the one I've been (very slowly) developing at- http://publiclab.org/notes/code4maine/08-05-2014/aerial-mapping-drone-for-under-60. Something like MultiWii would be easily modified since it is based on Arduino. Although the current iteration of my UAV project has abandoned the 8Bit Arduino in favor of a 32bit STM chip, it does include a cheap IMU called the GY-80, which includes a barometric pressure sensor, a digital compass and the MPU6050 10-DOF gyro/accel combo.

At this point however, I would strongly advise AGAINST using an STM over Arduino even if the 8Bit system severely limits capabilities... The enormous resources provided by the Arduino user-community outweighs the performance increase in ALMOST every situation I've yet tried...

Dan's response in the accompanying thread to this note in the discussion forum gave me an idea about retrieving the Glider in an ocean/large lake environment... What if the Glider could be configured to go into a "retrieval mode" where it surfaces and emits a locator signal via AIS? As readers may or may not be aware, I've been working on a low-cost method of receiving AIS signals at- http://publiclab.org/notes/ajawitz/06-11-2015/raspberry-pi-as-marine-traffic-radar. However, I know very little about what it takes to actually TRANSMIT... One thing I do know is there are many different classes of AIS transmissions assigned to different uses. So big commercial ships will use "Class A" transmissions, smaller vessels will use "Class B", Aids to Navigation use a dedicated class, search and rescue etc... AFAIK, one of these specifications has been set aside for "Scientific Research" but further information is hard to come by (chances are it has some kind of obscure acronym)... Anyway, I'm picturing something akin to the Slocum Glider where the tail of the vehicle can double as a flag/antenna in retrieval mode-

Or like this Glider from the University of Washington-

Challenges- AFAIK the SDR revolution only really extends to receiving radio signals and transmitting them is still quite complicated in terms of hardware and regulatory approvals... The closest thing to an SDR that can both receive and transmit is the HackRF but at over $300 this still hasn't quite hit the "affordable" mark...

Dan- Considering your background in both Glider research and amateur radio do you know anything about AIS specs for scientific research applications? If not can you recommend a good resource or place to start looking? Assuming the regulatory process itself doesn't prevent the development of a low-cost AIS transmitter for research purposes, are there any obvious reasons why such a device would be technically impossible?

I just can't resist the temptation to build one myself! Even though the DIY version is probably useless in an ocean environment, it just looks so tempting! Anyway, looking at the build instructions and the BOM one problem keeps popping up in that all the hardware components can only be ordered in bulk from McMaster-Carr. Seeing as you mostly only need one of each component (washers, nuts, bolts etc...) they only ship in packs of 100! I searched long and hard through Home Depot and Lowes and they don't seem to carry anything in those particular dimensions. The build instructions look pretty detailed, so its hard to tell which parts you can get away with using a different size. I guess I'll try printing out the printable parts first and then I can see what pieces need to fit.

One detail I really appreciate about the Seaglide project compared to something like OpenROV is the 3D Printed parts are mostly small enough to fit the build volume of the low-cost desktop 3D Printers (like the Printrbot SImple) that are finally entering the mass market. The only exception might be the "plunger" component which might be too tall for a desktop printers' Z-Axis...

This is a great question about bulk ordering. @mathew can you comment on this?

UPDATE 6/29/15:

I've managed to build most of the "buoyancy engine" with a large syringe, 3D printed parts, threaded drive bolts and metal "BBs" for ballast. I also got the right water bottle, and the Arduino Pro Mini.

Cross-referencing the BOM with inventories at local Home Depot etc... got me nowhere, so I just went ahead and printed out the 3D printable pieces of the buoyancy engine to serve as a reference point. Using this method I was able to find the drive bolt, brass nut and locking washer rather easily by fitting them into the printed space. The good news is the metal parts are relatively common. The bad news is none of my standard or micro size servos would fit the very specific dimensions of the printed parts. So now I have to wait until the servo with the proper dimensions arrives in a week...

The completed Buoyancy Engine with Servo Attached should look like this image from the official instructions

My initial impressions are that while very impressive, this is an unusually complex diy build in terms of the conceptual engineering involved. Its clear that it was designed by professional engineers as opposed to a casual tinkerer... At least as far as the buoyancy engine is concerned, this does mean that the materials used must closely match those of the original build as there is little room for alterations. This somewhat diminishes the DIY value of using common everyday objects when you really have to use the exact same water bottle, syringe, servo etc... Luckily, the parts are still relatively inexpensive and available online. I'm hoping the same rule does not apply to the "sensor pack" as this is where I would like to test some of the Open Water modules like RIFFLE.

One final observation... You can definitely tell this build came out of the Navy as more than any other DIY project I've ever worked with, I couldn't shake the feeling that I was building some kind of a bomb! Especially after loading the 3D printed cylinder with BBs for ballast and attaching what looked a lot like a triggering mechanism... Needless to say, I would avoid taking the completed glider through an airport...

I might have a temporary workaround for the control issue... At least until we can incorporate some kind of autonomous navigation... According to the instructions, the Glider should pretty much swim in a straight line. So if we need it to return, in theory it should only require one change of direction... A very low-tech solution would be to simply attach a long fishing reel to the front of the Glider. When the Glider is on its initial path, the reel will be set to slack and trail behind it. Once it reaches the end of the line the slack will tighten and the Glider will change direction. The line will then be carefully reeled back allowing the Glider to return on its own power. Alternately the line could be attached to the rear of the Glider but this would mean working against the its own propulsion motion. So rather than a "Thermal Fishing Bob" it would be more like a "Thermal Bobbing Fish"!!! You could even have some fun and pretend like you're fishing for robots!

Now that this project is taking on a life of its own, I figured it was time to give it, its own research note at- http://publiclab.org/notes/ajawitz/07-11-2015/buoyancy-driven-underwater-glider

Please do continue the discussion!