MIT AUV Lab

First Day of Deployment, 2010

2010-06-03T22:50:00.002-04:00

Here's our first day of deployments in 2010, as told in pictures:

Compass Error Calculation

2010-04-16T16:15:00.003-04:00

The results are in, and the compass did not do well. If the compass was reading correctly, then even with some level of error in each measurement we should see no trends.

These do not appear to be random errors. We will have to try this again with another INS and see if we get a similar result.

Compass Error Measurement

2010-04-14T16:09:00.001-04:00

Today we started investigating a possible error in our compass measurements by recording the INS heading and GPS coordinates. Basically, we mounted the INS and GPS on a piece of plywood, hooked up a laptop and car battery, and threw the whole setup into a non-ferrous plastic radio flyer wagon.

This is the path recorded by the GPS as we went around and across the football field. The goal was to get data in as many directions as possible, but long enough runs in each direction so that we'd get good averages.

Interactive AUV Mission Planning with Google Earth

2010-04-05T12:37:00.003-04:00

In winter, we develop.

For best results, watch the HD (720p) version in full screen.

The Search for Didemnum: Day 5 (Now THAT'S More Like It)

2009-10-15T19:26:00.003-04:00

We got off to a great start!

Clear water, proper bottom type, and great lighting. Today is a day to find didemnum!

With the vehicle fully functional after the previous day's repairs, Seth and Mike began fishing in earnest.

We're no experts on identifying tunicates, but we had no problem finding spectacular examples of sea life amongst the volumes of images that Odyssey brought back.

This is a sunflower sea star.

A skate...

A flounder, doing a good job of blending in...

A tiny solaster (literally, a sun star)...

After taking a few thousand pictures, we decided to seek out more rocky bottomed waters by moving further east.

Our recovery was complicated by the rocking of the boat, but as a team we are getting used to handling the swinging vehicle.

(There were also some good pictures of the water column!)

Our second deployment went very smoothly.

Unfortunately, the water was far more turbid in our second area of operation. Sponges, anemones, and redfish dominated the images.

The sponges look like potatoes in these images.

Although beautiful to look at, anemones are not what we are after.

As the light faded, we made our last recovery for the week and prepared to head back.

We had pictures of too many starfish to count, in all shapes and sizes.

This one was hanging out inside a mussel shell. I can't tell if it's eating or just resting.

There were also quite a few crabs, hermit crabs, and sunflower sea stars.

Divide and Conquer

2009-10-14T23:30:00.003-04:00

While Mike performed a miraculous teardown and rebuild of the vehicle on the deck of the Isabella & Ava,

Seth, Kyle, and I travelled back to Cambridge to put new oil into the thruster that had spent some time at the bottom of Gloucester harbor.

We met back up late that night.

The Search for Didemnum: Day 4 (One Good Mission, and Downhill From There)

2009-10-14T17:52:00.002-04:00

Today we were accompanied by Rachel from the Northeast Consortium (NEC), the organization responsible for funding this trip.

We decided to survey near the outflow pipe from Boston because we figured that it would be warmer and more likely to sustain didemnum.

Deployment was rough. We immediately lost the vehicle behind the stern and had it pop up near the starboard bow. That can't be good.

However, everything seemed to have survived intact, so we started our day of testing. After a few false starts, we ran a survey mission to see if the warmer water from the outflow pipe would make a difference in our images.

It did...

Whether from the pipe or from the gale force winds during the previous night,

...the turbidity in the water made for terrible image quality.

This lobster trap was one of the few recognizable objects in all our photos.

Seth had no luck finding any fish, and the sea state didn't help.

Jellyfish, on the other hand were found in abundance

throughout

the

water

column.

So

many

jellies!

We decided to haul up the vehicle to try another location. It was especially difficult with the waves being higher than they were yesterday.

Odyssey swung quite a bit while we guided it to its cradle.

That's when things got worse.

"We're all done, guys" said Seth.

"What?"

Besides the fact that the titanium thruster guard had been hit hard enough to crack one of its welds,

and the fact that the fiberglass CTD shroud had taken a beating,

it was the broken-off oil filling port in the thruster that got our attention.

We had been oeprating with slowly-leaking oil all day! Worse, we had no way to refill and re-close the hole; the show was over.

You can see the red paint that scraped off the bottom of the boat and onto our thruster.

There was no choice but to end the day early and attempt to do some emergency repairs.

Back to the lab!

On the way back to shore, John caught a bird that had found its way into the wheelhouse (miles from shore) and turned it loose outside.

Battery Recharge

2009-10-14T08:31:00.001-04:00

Our battery continues to have a much tighter difference between its max and min cell voltages than we did before we replaced supercell #10.

This is what a healthy battery looks like.

The Search for Didemnum: Day 3 (Survey Patterns and Pictures)

2009-10-13T22:58:00.003-04:00

Today we worked on our survey pattern behavior. We seem to have some troubles when heading due south.

Aside from that, the surveys look pretty good.

The images we are getting at 2m altitude are a little too pale. You can make out a scallop in this image, but not a lot of red channel is coming through.

We will have to fly lower to get good quality.

At 1.4 meters altitude, things are considerably better!

A crab.

A sea star.

A lobster.

A sunflower sea star, a sculpin, and a "regular" sea star.

We prepared to make the day's first recovery. Working as a team makes what would be an awkward (and somewhat hazardous) process go quite smoothly.

John (the first mate) is very good at coordinating the recovery process.

As the seas get rougher (from 25 kt gusts), deployment gets more "interesting". We try extra-hard to get the vehicle far from the boat before we let it go.

After putting Odyssey back in the water, we weren't able to accomplish much. There were some transient control issues in our way.

We did get a good sense of our minimum turning radius (at cruising speed) when we tried to hold a position.

All in all, we confirmed that the camera system is working reliably that our survey patterns actually look like survey patterns (even if dead reckoning is still a little behind).

Every trip to the surface has pictures filled with these jellyfish.

Ready to Re-deploy

2009-10-09T16:09:00.001-04:00

With the RTU overhaul complete, we are ready to get back to hunting Didemnum.

RTU Stress Testing (again)

2009-10-08T11:29:00.001-04:00

We did some more stress tests on the RTU after re-tightening the bolts.

Another RTU Overhaul

2009-10-07T12:21:00.000-04:00

Back at the lab, the RTU had to be removed so that we could service it.

After draining the oil and opening the housing, it was clear that some of the bolts holding the stator had worked themselves loose again. This is a known issue, caused by a shortcoming in the design of the RTU.

For now, the only thing we can do is re-tighten the bolts and re-assemble the unit.

The Search for Didemnum: Day 2 (Double Success, Double Failure)

2009-10-06T20:11:00.002-04:00

On our second day out, we went a few miles further offshore to try some surveys in roughly 100 feet of water. From this far out, we could see the Boston skyline relatively clearly.

The day's first deployment went smoothly.

The goals for our second day out were to find out what altitude is best for taking pictures, and to see if Justin's changes to the survey behavior could help straighten out the tracklines when the vehicle is at depth.

In this test, we stepped through altitudes from 1.8 meters to 1.1 meters in increments of 10 centimeters. The results are exactly what we had hoped for: consistently good images, and a better idea of what range of altitudes will produce them.

Notice the difference in the amount of red channel as the vehicle approaches the bottom. To get good information from red light, we have to fly very close to the bottom. A difference of a few tens of centimeters makes a big difference in light quality!

We also had one flash misfire, the cause of which is unknown.

Overall, this set of images is a huge success; we can start doing image surveys in earnest now that we know the physical focus and aperture settings are in the right place.

The new survey behavior also worked well... mostly. The tracklines are much straighter, but what happened at the 4th corner?

Before we could wonder too much, Justin noticed that the vehicle was running in circles under remote control and had a problem with rolling over.

Uh oh.

Mike and Seth sprang into action to check it out.

Both thruster cables were still attached. So, we decided that recovering Odyssey would be the only way to fully determine the cause of the problem. Kyle got to try his hand at recovery.

Once the vehicle was on deck, the problem was fairly evident. We had lost one of the ninja pins that holds the thruster in proper orientation on the RTU shaft, and the thruster had worked its way to the opposite direction.

Only 3/4" of the thruster was still within the coupling. Had it worked its way fully loose, we would have lost it permanently!

Unfortunately, there was a second problem as well: the RTU had fallen out of alignment, and was unable to produce adequate torque on the shaft to properly position the thrusters. This is a game ender; we have to take the RTU back to the lab for disassembly.

Catching a dogfish lifted our spirits somewhat. We threw it back.

On the steam back to the harbor, I was able to put together a small mosaic of a few good images using the hugin panorama stitcher (not to be confused with the Hugin AUV).

When the RTU is fixed, we should be in very good shape to get more like this!

The Proof: We Actually Did Fix The Battery

2009-10-06T09:03:00.001-04:00

It looks like despite the battery problems we had last week, our battery is healthy and doesn't have the one divergent supercell that we noticed during UNH testing.

You can also see the point at which we returned from dinner to find that the power supply had switched off.

The Search for Didemnum: Day 1

2009-10-05T21:45:00.004-04:00

Today we started our hunt for the invasive species Didemnum Vexillum. Our equipment was packed and ready to go

We left at 7am to beat the traffic on our way to Scituate.

Once we arrived, we spent the next 2 hours moving our stuff onto the F/V Isabella & Ava and getting it stowed properly.

The boat's crane was big enough (and the tide was high enough) to lift Odyssey IV out of the pickup truck and place it safely on deck.

REx II came too.

Mike, Seth, and Andrew re-assembled the vehicle: fins, thrusters, and thruster guards.

I pieced together the quick-release and recovery pole with appropriate hardware to attach to the boat's crane.

And away we go!

We didn't take any chances on our first deployment; we used the zodiac to pull the AUV away from the side of the boat. Better to err on the side of caution! Our first set of missions was now under way.

The first few dive missions were very disappointing -- the camera only able to capture overly dark images. It was clear that we would have to recover and adjust the aperature.

Rather than recover immediately, Justin took the opportunity to do some survey missions to test the vehicle's "survey behavior" functionality.

The first survey (done on the surface) was a poor performer. The vehicle didn't really stay on each of its tracklines before reaching the next of its waypoints.

Making the tracklines longer (again on the surface) produced a much cleaner lawnmower survey pattern.

Running the survey at depth showed us that we still have a few control bugs to work out. With that, it was time to recover the vehicle and adjust the camera.

Instead of taking the entire vehicle apart, we were able to simply remove the bottom fairing and open up the bottom end of the camera bottle with a strap wrench that they had on the boat. We were able to re-adjust the camera in record time!

Not shown: the recovery process. I got to try out my new recovery pole, and it worked like a charm! I snagged the vehicle on the first try.

With the camera adjustment done, it was time to re-deploy. We were much more confident the second time around.

If you look closely, you can see the horizon tilt as the boat heels over due to the weight of the AUV. This helped keep spacing between the AUV and the gunwale.

Our quick-release hook makes life very easy.

The next camera test was also disappointing.

"I think my eyes are getting better. Instead of a big dark blur, I see a big light blur."

These pictures were taken at a 2-meter altitude, and you can see hints of the sea floor if you squint.

We decided to go even lower -- to 1 meter -- which puts us at the limit of what our DVL is able to sense for altitude.

But we finally captured some good images of the sea floor with our new camera system! This is excellent!

The only problem is that the best image in this set (frame 236) was taken without a DVL lock, so we have no idea what altitude produced it. The images before and after it are no help either.

Since we had to get back to the harbor before low tide, it was time for the second recovery of the day.

If you look closely at the top row, you'll see the new recovery pole in action. The boat captain was an excellent crane operator and the vehicle was back in its cradle in just a few moments.

We steamed back to Scituate.

This was our one good result of the day. It's not much, but it proves that the camera's physical settings don't need to be changed; all we need to do is tweak parameters tomorrow to get more images like this one.

We set up the charger before dinner, but after eating we checked on it and noticed that the power supply had tripped. This was probably when the boat's power was switched over to the auxiliary generator.

Back to the hotel! One day of testing down, 4 to go!

New Recovery Pole

2009-10-02T15:48:00.001-04:00

I put together a recovery pole for Odyssey IV out of some spare steel tubing, some hose clamps, a carabiner, and a length of 3/4" EMT conduit.

The carabiner will be attached to the ship's crane, and will slip out of the pole once it has been clipped to the vehicle's lifting bail. We will find out how well it works next week!

Battery Scare in the 11th Hour!

2009-10-02T14:59:00.002-04:00

While we were re-installing the battery into its glass sphere (after replacing the under-performing cell #10), there was a soft pop, a crackle, and then a hiss as clouds of tan smoke filling the air.

Some part of the battery had short circuited; something was seriously wrong!

We had to take the battery apart again to see what happened and to try and fix it. We are especially under the gun to fix this, because we are supposed to deploy on Monday to look for didemnum! Yikes!

The source of the crackle became apparent; it was the sound of a wire sparking and melting itself out of its jacket. You can see the bits of melted black jacket on the red wire.

There were some burn marks on the outside of the battery too.

But, the real damage was on the corner of the cell where the smoke forced its way out.

We count ourselves very lucky that only one supercell was damaged and that nothing caught on fire. In the end, we had to use the supercell that had been removed as the replacement for the one that was damaged (although we recharged it). A net gain in the end, but only a small gain.

Around the Americas

2009-09-25T14:30:00.000-04:00

The folks on the Around the Americas cruise stopped by to build some sea perch and get a demo of our REx II vehicle.

Power Supply Crate

2009-09-14T11:36:00.003-04:00

One of the annoyances of transporting our support gear has been finding ways to carry and stack the power supply that we use to recharge Odyssey IV's battery. The charging cable is 50 feet or so and 1" in diameter -- a bit difficult to wrap your hands around when it's all coiled up. The power supply itself is large and flat with protruding pieces on all sides (rackmount ears, knobs, and terminals).

To kill 2 birds with one stone, I made a crate that uses the power supply for its bottom.

One of the constraints I put on this project was to be able to get the power supply out without using tools. So, I made some notches for the ears to fit into instead of using the rack-mounting holes. This keeps the power supply from sliding around and also keeps a gap around the edge so that air can get into its vents.

The front part was a bit more interesting. I had to be able to access the knobs as well as leave enough room to lift the power supply out of the notches. On the other hand, I had to prevent the cable from spilling out. The solution was to have a large wooden latch that would rest on top of the supply to retain the cable but could be flipped up to remove the power supply.

Visual Servoing at UNH, Day 3: Style points

2009-09-11T20:19:00.000-04:00

I had written a new web-based GUI for the vehicle's onboard database, with some open flash chart graphing capabilities. When we recorded the overnight recharging of the Odyssey IV battery, my code got its chance to shine.

As you can see, cell 10 seems to be defective and it will probably be replaced soon.

Since we were able to closed-loop control our position and yaw based on the camera, and our altitude based on the DVL, it made sense to close the loop on sway based on our heading.

This means that we can say "look at the south side of the pipe", and it maintains a position on the south side of the pipe. In this nav plot, we visited the east, south, west, and north sides.

We packed up and headed back to the lab. Another very successful trip!

Visual Servoing at UNH, Day 2

2009-09-10T19:53:00.004-04:00

The day started off very well, with us checking out the mounting pole for a piece of multibeam sonar equipment that was being tested earlier in the week.

Satisfied that the image processing component was working properly, we switched to the 13cm pipe that we would be using for our demo.

Almost immediately, we ran into trouble... all of our calculations for distance were off. We ran and re-ran our calculations for the angular resolution of the camera, and for calculating our distance from the pipe based on its width in pixels on the video.

An hour or 2 later, we noticed that we had accidentally declared our angular resolution as a boolean instead of a double. That's a pretty big error! Surprisingly, it only affected our distance calculation by a factor of ten. But once we fixed it, we were spot on.

The results are dramatic!

Visual Servoing at UNH, Day 1

2009-09-09T18:07:00.001-04:00

Today we are at UNH to work on some visual servoing code that Justin wrote. The goal is to be able to follow a vertical pipe, controlling yaw and position based on image processing.

Mike has mounted the camera on the bottom of Odyssey, facing forward so that we will be able to track an object in front of the vehicle.

Early on, we had to re-adjust the aperature on the camera. We had made it as small as possible to increase our depth of field, but it was just too dark in the facility to see anything on that setting. So, we opened up the aperature and double checked the focus.

With the vehicle on an ethernet tether, we were able to use coriander to tweak the settings of the camera once it was underwater. This program makes the tuning stage go very quickly!

REx II is Back in Cambridge

2009-09-03T15:03:00.000-04:00

Today, we got the REx II vehicle back from Hawaii.

Due to restrictions on shipping, we didn't bother to send the lithium ion battery by mail. Instead, we just built a new one.

Unlike the battery for Odyssey IV, this battery is designed to fit into the bottom half of a pressure tube. The individual supercells are not wrapped in plastic, but instead squished together to form the full battery.

The electronics fit on top.

Potting: A Mini-Howto

2009-09-01T11:20:00.002-04:00

I put a mini-howto on potting on our wiki.

Nantucket Sound Testing, Day 5

2009-08-21T20:36:00.001-04:00

Today we did some tests to characterize the effects of our horizontal stabilizing fins on our depth while under way. The goal was to see if we could fly the the vehicle like a plane, changing altitude by changing forward velocity. As it turns out, there isn't a huge effect, even when we pointed them down significantly.

The final recovery worked like a swiss watch.

Taking Odyssey off the boat when we got back was also a very smooth operation.

So long, Mike and Drew!

All in all, a very productive trip.

Nantucket Sound Testing, Day 4

2009-08-20T17:09:00.002-04:00

Today some grad students from CSAIL came out to try running some of their planning code on Odyssey.

It was a calm day, so -- combined with the fact that we are getting better at the deployment process -- everything went very smoothly.

One of the students went for a swim, so Justin gave her a tow on the vehicle.

We weren't the only ones towing thrill-seekers around the cove...

We did a lot of dive missions to test out the camera system. The vehicle makes a splash on the surface, then pops up a few minutes later (and our confidence in this process has grown considerably since the beginning of the week). Unfortunately, it looks we will have to focus the camera manually since every picture between 1.5 and 5.5 meters altitude comes out blurry.

This time, Mike Ryan picked up a recovery pole from the WHOI facility, which makes it much easier to clip the crane hook onto the lifting bail. My butt didn't get wet this time, and I was nowhere near falling off the dive ladder.

The rest of recovery is a snap.

Nantucket Sound Testing, Day 3

2009-08-19T17:58:00.002-04:00

First thing this morning, we went to get our DSSI thruster fixed at their facility in Falmouth. As we had hoped, there was no damage to the thruster, so they put a new cable on it and we were good to go.

We ran almost nonstop throughout the day, setting up shop on the fly deck so we could get a better view of the vehicle.

That's it, off in the distance.

We made a lot of progress on tuning the PIDs for yaw, although the gains that produce some of our straightest trajectories also produce some of our most spectacular instabilities.

When the mechanical engineers are bored, it means the software engineers are making good progress.

Nantucket Sound Testing, Day 2

2009-08-18T20:16:00.004-04:00

In the morning, the crane operator moved Odyssey onto the boat -- it was a matter of minutes. Life is easy with a giant crane truck.

The crane on the boat was also really convenient for putting Odyssey in the water.

Taking the vehicle out of the water was also fairly painless, which is good because we had to do this a few times while adjusting the trim.

Mike operates the crane, Mike and Seth fend the vehicle off of the gunwales, and Noah guides both the crane and the quick release. We're still getting the hang of this part.

Noah is about to pop the quick release.

As the vehicle pops free, we have to work to keep it off the side of the boat while Justin takes the controls.

We do a good bit of manual control to maneuver the vehicle in tight situations, and to get it set up for missions. This will get a little easier as we get the control system and behaviors set up.

While Justin and I work on some simple missions to test the control system (goal: good yaw control without vertical stabilizing fins), Seth tests whether fish are attracted to bits of hot dog.

Then something familiar happens...

The vehicle starts going in circles when it should be going straight. Have we lost a thruster again? Seaweed? Mike and Seth check it out in the zodiac... apparently the cable got wrapped around the thruster and had disconnected.

So, Seth grabbed a handful of zip ties and went to secure the cable.

Unfortunately, the thruster cable had been damaged internally... so the day was over early. Meanwhile, Mike Ryan and Noah had upgraded the refrigerator on board to a model with a freezer. We debated putting an outboard on it and having a boat race.

But the day wasn't a total loss! Seth caught a fish with the hot dog after all, which he filleted and stuck in the new refrigerator. Possibly for lunch tomorrow!

Nantucket Sound Testing, Day 1

2009-08-17T21:52:00.001-04:00

We've arrived in Falmouth, MA at the Woods Hole pier, to meet up with Ryan Marine on the research vessel Discovery. The crane operator had left for the day, so Odyssey will spend the first night on the pier.

This vessel is a bit larger than the M/V R&R that we used previously. Our support gear fits comfortably on deck and is reasonably protected from the elements.

Night Testing: Camera Tests

2009-08-14T11:37:00.000-04:00

At our second night at UNH, we planned ahead and bought plenty of food before the stores all closed. It helped keep us awake to take food breaks throughout the night, but we all took periodic naps when we could.

Although the camera driver and strobe system worked, the images were a little too bright (and slightly out of focus, even at long distances). We will have to adjust the focus and the aperture before doing any survey work.

At dawn, we packed up and left.

Night Testing at UNH

2009-08-13T09:07:00.001-04:00

To tune up our new camera and flash system, we planned to take a trip up to UNH to use the large pool at their Ocean Engineering facility.

They were already booked for the week we wanted. But because we had a fairly urgent need to get our testing done, they agreed to let us use the pool "from 7pm until dawn". Time for some all-nighters! We loaded up the truck in the afternoon instead of the evening.

The UNH facility is more than large enough for our needs. It would be great to have access to a crane like this at a pier somewhere... or in our lab, or in an MIT pool facility.

It really makes testing more productive when you can get the vehicle in and out of the water quickly!

Our major accomplishment for the day (night) was getting the PID gains tuned for depth/altitude control. They are tight to within 1cm of the target altitude!

The one nice thing about working at night is that you can turn off the lights to see how effectively the strobe system illuminates the bottom.

Field Kits

2009-08-10T14:55:00.003-04:00

Today I organized some field kits to make deployments go more smoothly.

With the exception of those rare occasions where nothing goes wrong, bringing the right amount of stuff and having it organized -- so that you can find what you need when you need it -- is key to success.

We Need a Bigger Boat

2009-08-06T16:43:00.000-04:00

Pulling a 1000lb vehicle with an inflatable boat is difficult...

And in the middle of a shipping lane, it's a nail-biting experience.

Back On Track

2009-08-05T18:35:00.000-04:00

With a fixed thruster arm, we are back to tuning up the control system. Testing went well, but there was an extremely tense moment when the vehicle started turning in circles when it should have been moving in a straight line...

We thought we had lost a thruster again.

Fortunately, it turned out to be some seaweed that had been floating on the surface, which got tangled in the thruster and severely reduced its efficiency.

Damage Report

2009-07-30T09:41:00.001-04:00

We took a look at how our 3D-printed thruster coupling failed.

You can see in this picture that the coupling lost a big chunk on the side opposite from the direction of thrust. Basically, it cracked under the force of the thruster alone.

The amount of empty space in this printed part is clear now, and it's an open question whether the failure was related to salt or rust buildup inside these spaces.

There is mild corrosion on other parts of the thruster, about what you'd expect for 2 weeks underwater

Against All Odds!

2009-07-29T04:33:00.002-04:00

Today was a great day!

Two divers helped us find the lost thruster, after 2 weeks of being lost.

We got special permission from the assistant Harbor Master to conduct our search in the middle of the shipping lane.

"Please tell me this thing's expensive," he said.

"About $15,000".

He nodded.

The divers searched by taking 2 compass readings and heading out to where the lines crossed. They worked by dropping an anchor with their dive flag, and making concentric circles around it, getting further and further out as they searched in the muck by hand. After a while, they found... nothing.

We turned back to the Harbor Master. Since we had a GPS fix of where the vehicle was when it started going in circles, we figured we should use that technology instead to compasses to find the start of the search pattern.

"Let us do just one more try, and if this doesn't work we'll give up for good".

He said OK.

As it turned out, the new center of the search pattern was pretty far from where the compass readings had put it.

So the diver started again. It seemed like they were down for much longer on the second search, and we realized that we had probably missed again.

But when they came back up, they had a thruster raised above their heads! SUCCESS! They had found it on the next-to-last circle that they planned to make, and it was buried under 6 inches of sediment. The diver had been pawing through the muck, and hit something solid: OUR THRUSTER!

Take a good look at these guys, they are heroes!

Finished the New Thruster Arm

2009-07-23T15:16:00.006-04:00

The 2 machined pieces.

Fitting and comparing to the existing arm... our version looks as good as the one that was machined professionally!

Seth prepares to weld it up

ZAP, CRACKLE

Roughly filed down, this part is complete! Now we just need to have it annodized.

Thruster Arm Progress

2009-07-22T20:02:00.003-04:00

The thruster arm's end is nearly complete.

Plan B: A New Thruster

2009-07-21T16:34:00.003-04:00

Since we have had no luck finding our lost thruster, we are moving ahead with "Plan B", machining a new thruster arm to hold our spare thruster. We will continue the search for the original, but at least now we can plan to be back in the water and testing by a specific date.

Since this is fairly simple fabrication, it is far faster and cheaper for us to do the machining ourselves at a local shop.

An Unsuccessful Search

2009-07-15T17:35:00.000-04:00

Today we tried (unsuccessfully) to recover the thruster. We lost it in the middle of a shipping lane in very turbid water. Visibility is poor, and the panning underwater camera setup we are using has neither the field of view nor the accuracy to properly search for a lost object.

Our zodiac is probably not the ideal support boat for this operation either.

Disaster Strikes!

2009-07-14T20:10:00.001-04:00

Our 3-D printed mechanical fuse for the RTU coupling cracked and broke, taking our thruster to the bottom of Gloucester harbor.

Justin: Mike, the vehicle is going in circles instead of forward. Can you check the connection to the thruster?
Mike: Sure.
(Pause)
Mike: What thruster?

It would have ordinarily been retained by the electrical cable but ... we had just set it up to unplug rather than put strain on the other cable. "Whoops" doesn't quite capture the disappointment.

Bypassing Sliprings

2009-07-10T11:53:00.001-04:00

Of all the users of Parker DriveBloks, we seem to be the only ones that experience frequent capacitor failures. Of the failures we have, only the thrusters that go through the RTU (and its sliprings) seem to suffer.

Solution: bypass the sliprings by drilling a hole in the fairing for the cable. As a safety precaution against continuous rotation of the RTU (disabled in software, but... you never know), we aren't using the plastic shells to hold the cables together.

There's Your Problem...

2009-07-09T15:21:00.003-04:00

This image fails to accurately capture the cause of our non-functioning camera: the grey and purple wires being crushed in the threads of the housing.

Since this was largely the result of wires and plugs being crowded near the top of the bottle, I drilled a hole in the mounting plate for the PC boards and routed some of the wires to the back side.

First Day of Field Operations 2009

2009-06-30T18:30:00.001-04:00

Today was our first day in the water after installing the new flotation sphere (and a few other components) into the vehicle. The goal for the day was to properly trim the vehicle -- get its weight and buoyancy evenly distributed fore and aft, port and starboard. Our secondary goal was to get comfortable with the process of deploying: loading and unloading the truck, bringing the appropriate equipment, and so on.

In this picture, we've just finished unloading Odyssey IV from the back of the rented box truck (using the lift gate) and are assembling it next to the pier.

The Gloucester Maritime Heritage Center is a great place for us to test, because of their easily-accessible crane that is poised over water deep enough to hold the vehicle.

Only the height of the hook is controlled, so we have to use a lot of ropes to keep the arm of the crane pointed in the right direction and the vehicle from spinning.

After a few trips into and out of the water, we got the right mix of weight and buoyancy in the front and back. Time to haul it out of the water one last time and put all the buoyancy in the inside. The sun even came out!

We put the vehicle on a tether and did some joystick-controlled thruster tests. Manual control for this vehicle is done on the equivalent of a USB Playstation controller, although the available commands don't translate perfectly to the analog joysticks on it.

The thrusters are powerful enough to make visible effects on the surface of the water, and to tow the Boston Whaler that we were using as a chase boat.

Installation of New Battery

2009-06-25T15:00:00.000-04:00

The new battery has been constructed, and packed into its protective Nautilus sphere housing.

It appears to charge and discharge normally!

Acoustic Modem Road Case

2009-06-24T15:46:00.000-04:00

To make it easy to lug the acoustic modem, deck unit, and sound baffle into the field, Roman and I cut up pieces of foam to safely (and efficiently) pack those pieces into a shipping crate.

We cut some rough shapes into the green foam for the smaller pieces of the sound baffle so that it would be clear if any were missing.

Not shown are the deck unit (shipped back for repair) and the lead weights which will be the ballast for the sound baffle (we don't yet know how many we need).

RTU Shaft Stress Test

2009-06-24T13:57:00.001-04:00

To make sure that the force produced by our thrusters would not bend (and bind) the RTU shaft, we hung 60lbs of weight from each arm (simulating 30lbs of thrust at the end of the arm).

The RTU survived without binding!

Tool Holders

2009-06-23T10:15:00.000-04:00

To make it a little easier to work on Odyssey, I put together some tool holders that clamp onto its frame.

The construction is simple: bolt together a magnetic tool strip, angle brackets, spring clamps, and rubber feet.

This is what they look like in action.

Battery Reconstruction Progress: 50%

2009-06-16T09:45:00.001-04:00

12 of the replacement supercells (for our 24-supercell battery) have been completed. This time, we made better notes of the process.

Update: there is a new page on our wiki for the construction of a supercell.

Teledyne Benthos Acoustic Modems Have Arrived

2009-06-09T13:10:00.000-04:00

The deck box and top-side acoustic transducer.

The transducer that will be mounted on Odyssey IV.

Odyssey IV on Display after MIT Commencement

2009-06-05T16:34:00.000-04:00

Today we brought Odyssey IV to the lawn in front of Building 14 (Hayden Memorial Library).

We got a lot of great questions from graduates and their families.

Sea Perch Was a Hit at Maker Faire

2009-05-31T22:53:00.000-04:00

Our booth at Maker Faire was a big hit!

People of all ages (although mostly kids) flocked to the Sea Perch controls for a chance to drive the ROVs around the small pool we brought. Everyone was smiling!

2 kids built their own Sea Perch in our booth in a single day. Great job!

And to top it all off, we won 2 Editor's Choice blue ribbons! Thanks to everyone who stopped by to check us out, and please check out the plans on our Sea Perch website to build your own -- Sea Perch is an open source project!

Final Preparation for Maker Faire

2009-05-26T18:45:00.002-04:00

Today we made one last push to make more sea perch kits for Maker Faire. We fly out on Thursday, and start setting up on Friday!

We made nearly 90 kits in one day -- our most productive day yet!

We also finished the joystick panels using wood that we recovered from the disassembled roller coaster on east campus.

Sea Perch Kitbuilding

2009-05-21T14:26:00.000-04:00

We are rushing to finish building all our sea perch kits before they must be shipped.

Sea Perch Joysticks

2009-05-20T15:18:00.000-04:00

While the control boxes on a normal sea perch are designed to be economical, our demo at Maker Faire will need to handle 2 days of heavy use by (hopefully) excited people. We wired up a controller using arcade-style joysticks and mounted it to a sheet of plywood left over from packing ORCA in its crate.

ORCA V Packed and Ready

2009-05-19T16:14:00.001-04:00

We packed ORCA V into its crate for delivery to Maker Faire, using plywood, ratchet straps,and foam board.

Flotation Spheres

2009-05-14T11:51:00.000-04:00

Our 17" flotation spheres from Teledyne Benthos arrived today. They should help us squeeze a little more buoyancy into the AUV.

ORCA V's Final Journey

2009-05-13T17:19:00.003-04:00

Today we reassembled ORCA V, in preparation for its exhibition at the Maker Faire. After that, it will be sent to the MIT Museum for posterity.

DC1394 Camera Driver for MOOS

2009-05-11T17:47:00.004-04:00

Here is the first set of test images from the new DC1394 camera driver (and new generalized camera framework) in MOOS.

The driver enumerates its implementation-specific settings as individual MOOS variables to make administration (through uMS) easier.

In this case, the brightness setting was adjusted -- arbitrarily -- from its 32 (the default) up to 127 and down to 1.

Preparing Sea Perch Kits for Maker Faire

2009-05-08T16:55:00.000-04:00

In preparation for our trip to the Maker Faire at the end of May, we are putting together 350 Sea Perch kits.

The hardest part so far was cutting 14,000 feet of Cat 5e cable (network cable) in to 350 pieces and coiling it. The blue cables on the floor here are about 200 of those.

Dead Battery

2009-05-04T13:53:00.001-04:00

Our battery has failed.

After 3 years of faithful service, our battery was accidentally allowed drain far below the acceptable minimum voltage for Li-ion cells, effectively ruining its capacity. In fact, it would be unsafe to recharge this battery in its present under-voltage state.

We will have to build a new battery. This is unfortunate, but since these cells lose 20% of their capacity per year in perfect storage conditions, it was inevitable.

This will push our schedule back a bit...

Getting the Hang of 3D Printing

2009-05-01T16:30:00.000-04:00

This is a scale model of the Odyssey IV vehicle, made on our 3D printer. It was supposed to help us test the design of the recovery cage that we printed out, but we made the walls too thin... so instead, it will probably become a mobile in someone's office.

The Enemy

2009-05-01T09:44:00.000-04:00

The white gooey-looking mass covering this mussel is called "didemnum", an invasive species of sea squirt that's taking over the sea bottom in various locations off the east coast.

Although it looks like mucus, this animal is fairly firm, almost leathery. It's actually a large group of sea squirts -- cloned from a single individual -- tightly packed together.

We will be running some survey missions between mid-July and mid-August to learn more about where this species lives and how it is spreading.

Found a Good Boat

2009-04-29T16:21:00.002-04:00

Today we checked out a fishing vessel in Scituate that looks like it will make an excellent base of operations for our trip in August surveying didemnum growth.

The boom has great reach and can hoist 5000 lbs!

A few new features to pHelmSG

2009-04-29T15:06:00.008-04:00

One of the most ~~annoying~~ dangerous things about operating our vehicle is having typos in the mission script. When coding up the string of behaviours that are collectively a mission, an inadvertent typo or even just an oversight can lead to the mission ending too soon, critical behaviours not getting called, etc. There has never been any way for the user to get any kind of feedback of the text-based mission script, other than checking it twice... & then a third time.

To remedy this, I added a mission preview feature. The missions now generate a text-based mission graph & transmit it (through a MOOS variable called MISSION_CHART) to the GUI app running on the operator's console. The GUI then generates a preview image using the GraphViz, and displays it for the operator... before they hit the "START" button.

Here's an example:

In this example, the AUV dives down to 4.5 meters of alititude, then maintains 2.5, then 2.0, then 1.5 meters, before surfacing. You'll notice the various safety behaviours running in addition to the main mission, as well.

There are certainly some improvements to be made. I'd like the flag-nodes in the chart to be much less prominent than the behaviour-nodes. Also, the behaviours-nodes should give some more description of the behaviour's attributes. The hooks are already in the behaviour-base class to do this... I just need to write it into the child behaviours themselves.

In looking at the above example, you may have noticed (if you are at all familiar with looking at MOOS scripts), that the various behaviours now have more than one way of exiting. Successful completion, timing out, & error conditions all throw different flags. I've often had a behaviour timeout, or worse, fail & then the mission proceeds to the next behaviour as if it were successful. I'd prefer to have a little more control over the program flow when a behaviour is running.

Coming soon: loopable behaviours.

It Went "POP" and Died

2009-04-27T16:56:00.006-04:00

On Friday afternoon, while preparing for this week's in-water testing, there was a loud "POP", and a few seconds later the benchtop power supply's current-limit light turned on.

Uh oh.

Today we went about troubleshooting it. When we opened the main pressure housing, the smell of burnt electronics was noticeable (it soon dissipated).

The current limit on the benchtop power supply was lower than that of some of the fuses on the vehicle, so it was a simple matter of taking fuses out one by one to determine which component was shorting. It turned out to be one of the Parker Driveblok units.

The most likely explanation for the short was a failed capacitor. The Drivebloks have replaceable capacitor banks, so that was the first thing we checked.

This is the old capacitor bank, and the failed capacitor is quite noticeable. Not only does it have a considerable bulge overall, there is one spot with a prominent outward dent.

The reason for the failure is currently unknown. However, we did notice that the water blocks between each pair of Drivebloks were vapor locked -- the coolant enters and leaves on the bottom of each water block, and the top part had filled with air -- preventing any coolant from flowing. The rise in temperature may have contributed to the failure.

After putting in one of our spare capacitor banks, the system seems functional again.

Finding a Test Site

2009-04-24T15:55:00.001-04:00

Today we checked out the Gloucester Maritime Heritage Center as a possible place to do basic vehicle testing. After talking with the nice folks who work there (as well as the Harbormaster and Coast Guard), we are all set to go next week! There is plenty of room on the pier for us to set up a tent with all of our stuff.

There is also a good crane for lifting the AUV in and out of the water.

The pier is fairly long, and although we won't be doing much work in this narrow patch of water, it's a convenient place to trim and do basic tests before we turn it loose in a larger area.

3D-printed Recovery Cage

2009-04-23T17:44:00.001-04:00

Here's a timelapse video of the 3D printing process for a model of a vehicle recovery system.

This is the finished product with the support material dissolved away.

The design is meant to guide the vehicle into a solid structure that can survive some rough handling as it is pulled up onto the deck of a support ship.

RTU Stress Testing

2009-04-15T14:21:00.003-04:00

This is a video of our RTU doing some basic movements. Previously, we had a problem with the high torque of the motor putting enough strain on its fasteners to work them loose. So, after some adjustments to the motor controller (clamping maximum acceleration, adjusting damping, etc), it's time to find out how well they work.

Basically, we just rotate the RTU motor back and forth between two positions for a half hour.

The tapping sound you hear is coming from the hard stop that the motor makes. It took a bit of adjusting to prevent ringing -- that is, to prevent the momentum of the thrusters from causing the RTU motor to overshoot and wiggle to a stop.

The plastic bags covering the thrusters are keeping the pressure-compensating oil from slowly dripping onto the floor, and the thrusters are mounted in opposing directions for balance; normally, they would face the same way.

Time-lapse video of a 3D printing job

2009-04-10T11:35:00.001-04:00

This is what a 7 hour build looks like on our new 3D printer (filmed at 1 frame per minute).

Coolant Reservoir

2009-04-08T09:59:00.000-04:00

We added a reservoir to the vehicle's water cooling system to help keep air bubbles out of the lines. This should also act as a good temperature buffer when the motor controller boards have to do some intense work.

Our New 3D Printer Has Arrived

2009-04-07T19:34:00.000-04:00

Today, the lab got a 3D printer delivered (and assembled)!

This is a fairly amazing machine -- able to turn computer-drawn models into physical objects in ABS plastic! What makes this significant is the fact that ABS plastic is strong enough to be used in production parts, not just prototypes or mock-ups.

The most likely use for this machine will be to print "mechanical fuses" -- parts that are designed to fail early in order to protect more complex or expensive parts -- because it will make it easy to fabricate identical replacement fuses.

The printer works by extruding a very thin (0.01 inch) bead of plastic out of a heated tip. Layer by layer, the printed item is built up.

You can see the criss-cross pattern here, created from the first 2 layers of white plastic. The black plastic underneath is for support, and is removed (dissolved in a caustic bath) after the printing is complete.

Visitors from Essex Aggie!

2009-04-03T15:22:00.003-04:00

Today, Ann Witzig and some of the junior class from Essex Aggie stopped by our lab to learn about what we do and how to make vehicles of their own. Currently, they have built 2 of our sea perch ROVs, and were interested in ways to take it even further -- building their own AUVs/ROVs, and adding new components to the sea perch. Great ideas!

We were excited to show these kids all the cool stuff that our vehicle does... maybe a little too excited, since it's rare that people come in to check it out all the hard work that we do.

Hopefully they'll stay in touch! We'll need a new crop of vehicle engineers in a few years.

Highlights of Rex2 Footage

2009-03-31T11:08:00.001-04:00

Dylan is working with the Reef Explorer 2 (Rex2) in Hawaii. He just posted this video of some of the best recorded footage.

There were quite a few sea turtles curious enough to swim up to the vehicle.

Image Capture from a Basler A102fc in Linux

2009-03-30T15:10:00.000-04:00

Even though this image is nothing to write home about (by photography standards), it's one of the first I was able to capture using a Basler A102fc -- a firewire camera that we use in our vehicle.

Capturing this image in Windows is fairly straightforward; they provide a driver disc and some sample code. In linux, you need the libdc1394 library and some patience.

Before even getting to the point where I was worried about software, the first hurdle was how to power it on. The camera gets power from its 6-pin firewire port (4 pins for data, 2 for power) but my laptop has only a 4-pin port (no power).

The first idea was to pick up a powered firewire hub from the local electronics store. However, this led to a lot of frustration on my part before discovering that a powered firewire hub does not ... actually ... power ... firewire. It just amplifies the data signal so it can be repeated on the other ports!

One cannibalized firewire cable and 9V power supply later, I had my own powered firewire hub.

The software took only slightly less time, as the documentation for the libdc1394 library covers versions 1.x while the library itself is up to the non-compatible 2.x. However, I made great use of an existing project called grabdma.tgz.

After finding the details of the video mode I wanted using coriander, I set the proper image conversion function (the camera returns YUV images, not a Bayer pattern) and everything fell into place.

Opening the Main Pressure Housing

2009-03-18T11:00:00.001-04:00

The majority of the electronics inside the Odyssey IV are inside a glass sphere near the top of the vehicle's interior. Today we opened the sphere to do some work calibrating the RTU.

There is quite a lot of stuff packed into this 17" diameter space, including the PC104 mainboard, 2 large thruster motor controllers, and the RTU controller. The water cooling system you see here is actually for the thruster controllers, not the main CPU.

With the top half of the sphere removed, you can see the method we use to quickly disconnect the inside of the sphere's penetrators from the rest of the electronics.

New Hires

2009-03-03T17:30:00.003-05:00

I'm one of two new hires to the MIT AUV Lab (along with Seth Newburg).

This lab is where our vehicle lives when it's not in the water.