Whats better than 1 AUV? 2 AUVs!

A spot for our autonomous cousins.
Post Reply
Seawolf
Posts: 33
Joined: Oct 15th, 2011, 10:22 pm

Whats better than 1 AUV? 2 AUVs!

Post by Seawolf »

Hey,
I'm a grad student working on a collaborative AUV project. We've just started ordering some parts, so hopefully there will be something to show off in a couple of weeks.

Our basic system design looks something like this (pardon the MS paint):
thePlan.png
thePlan.png (25.47 KiB) Viewed 11079 times
The current plan is to have each underwater vehicle tow a wireless router in a buoy. Each vehicle will have a network camera connected to the wireless router, sending video footage back to a base station for analysis. The wireless router will also be connected to an arduino on board the vehicle. The arduino will read sensor measurements and relay them back to the base station. It will also recieve commands from the base station and drive the thrusters accordingly. Much of this architecture is based off of the wifi rc car found here.

For thrusters, we're planning on having 2 thrusters for forward/backward and turning, and another 2 thrusters for depth and pitch. Our initial design looks something like this:
cad12.png
cad12.png (104.46 KiB) Viewed 11079 times
(PS: I know the thrusters on top don't actually provide thrust in that configuration, this was the marketing picture :twisted: )
We're planning on 1000GPH bilge pump cartridges all around, powered by an onboard SLA battery.

I'll try to keep this thread updates as we make progress.

Cheers,
-Seawolf
Alaska Rov Systems
Posts: 69
Joined: Dec 6th, 2010, 6:24 pm

Re: Whats better than 1 AUV? 2 AUVs!

Post by Alaska Rov Systems »

Nice design, I am planning on building one that has an on board rc rx unit and have the antenna go to the surface to a floating buoy. I don't want a tether on the auv/rov. i want a complete wireless system with onboard batteries, lights and thrusters. I figure that the radio signals should be able to hit the antenna on the surface and then travel to the rx unit and turn on the thrusters.
As a grad student do you have any suggestions for me on either designs, onboard power systems or if rc wireless control will work as i have stated above??

thanks,

Alex
Alaska Rov Systems
Seawolf
Posts: 33
Joined: Oct 15th, 2011, 10:22 pm

Re: Whats better than 1 AUV? 2 AUVs!

Post by Seawolf »

The battery choice will depend on how much current your thrusters draw, and how long you want to operate.
I'm going with one of these per robot for power, because its the largest capacity battery I can find that will fit in a 6" diameter pvc pipe. Based on the spec sheet, that should allow a vehicle to run 2 thrusters for about half an hour constantly.

I don't think the antenna stretched to the surface would work, just because of interference from the thrusters. Motors put out a lot of electromagnetic noise, and RC signal strength isn't great to begin with. Your best option would probably be to have some type of receiver on the surface of the water which can amplify the signal and send it down to the vehicle.
Seawolf
Posts: 33
Joined: Oct 15th, 2011, 10:22 pm

Re: Whats better than 1 AUV? 2 AUVs!

Post by Seawolf »

Progress Update
We've been putting a lot of effort into designing stuff, which is boring to talk about. But we're finally to the point that we can do some fun stuff.

This past week we made a good bit of progress on the controls. We showed the ability to transmit images over a wireless network using a network camera, and the ability to process these images and send commands to a microcontroller based on the images. The setup looks something like this:
prog1.png
prog1.png (6.44 KiB) Viewed 11003 times
To show that it works, we made a program that stops the motor when the camera sees a red object. Our autonomous search function will be based on this framework.

Here's a screenshot showing the color detection.
Lab3edit.png
Lab3edit.png (310.38 KiB) Viewed 11002 times
In the next couple of weeks we plan on doing some thruster tests, and modify a wireless router to allow us to send commands to the microcontroller over a wireless interface. The outcome of the thruster test should tell us enough that we can buy stuff and start building.

Cheers,
-Seawolf
Alaska Rov Systems
Posts: 69
Joined: Dec 6th, 2010, 6:24 pm

Re: Whats better than 1 AUV? 2 AUVs!

Post by Alaska Rov Systems »

Thanks for the info!

Alex
Seawolf
Posts: 33
Joined: Oct 15th, 2011, 10:22 pm

Re: Whats better than 1 AUV? 2 AUVs!

Post by Seawolf »

I'm still alive. Once finals are over I'll be able to do some real progress updates. But for now, an updated CAD model and a quick video will have to do.

Here's what each vehicle will look like. There's a yardstick above the robot for reference.
Robot2.jpg
Robot2.jpg (223.16 KiB) Viewed 10874 times
And here's a view that shows two robots collaboratively moving a 10' long PVC pipe.
0CradlePool2.png
0CradlePool2.png (172.79 KiB) Viewed 10874 times

Finally, here's a video of our progress this semester. We've built a robotic boat that can be remote controlled from a laptop. It can use its onboard sensors to travel in a (fairly) straight line over reasonable distances, and it can use a camera to track colored objects. Next semester we're going to be stuffing all the electronics into a pressure vessel and adding more thrusters to make it an underwater robot.
Seawolf
Posts: 33
Joined: Oct 15th, 2011, 10:22 pm

Re: Whats better than 1 AUV? 2 AUVs!

Post by Seawolf »

And... I'm back! In the past couple weeks we've gutted our boat and started turning it into a AUV. We've got a hull made of 6" PVC pipe with a camera dome in the front. (Shown below)
hull
hull
0131121435b.jpg (189.74 KiB) Viewed 10759 times
It needs to be cleaned a little, but you get the concept. We've also got a mockup of the frame that will allow us to mount thrusters, ballasts and weights.
frame
frame
0131121435a.jpg (175.22 KiB) Viewed 10759 times
We came up with a pretty neat system of hose clamps and pipe straps that allow us to easily move the motors around so we can find the optimal position for them. I highly recommend it.

Finally, we have all our electronics wired up and mounted. Note: our camera isn't shown, we're still mounting it. The IMU is also not attached to the board yet, because its expensive and we don't want to damage it.
0131121437.jpg
0131121437.jpg (230.47 KiB) Viewed 10759 times
0131121437a.jpg
0131121437a.jpg (170.66 KiB) Viewed 10759 times
We decided to try having the wireless router inside the robot, with a long antenna wire trailing to the surface. We haven't tested this part yet, so we have no idea if it will work.
0131121438.jpg
0131121438.jpg (201.79 KiB) Viewed 10759 times
We're getting pretty close to having a robot in the water. We need to do some more testing of the pressure vessel, because it would be rather expensive if it failed with the electronics onboard. We're also in the process of finding a distance sensor that works well underwater. After failing to waterproof a hobby ultrasonic sensor, we've ordered one of these http://www.ebay.com/itm/Waterproof-Ultr ... 0721354139. Hopefully it can be made waterproof, so that our robot can find its way around the pool on its own.

One last teaser about how the final robot will look:
0131121435e.jpg
0131121435e.jpg (195.6 KiB) Viewed 10759 times

Cheers,
-Seawolf
Alaska Rov Systems
Posts: 69
Joined: Dec 6th, 2010, 6:24 pm

Re: Whats better than 1 AUV? 2 AUVs!

Post by Alaska Rov Systems »

I like that youtube video ! Very creative at the beginning with the jaws music!


Alaska Rov Systems
Seawolf
Posts: 33
Joined: Oct 15th, 2011, 10:22 pm

Re: Whats better than 1 AUV? 2 AUVs!

Post by Seawolf »

Well, its been a while. Here's a quick update and some picture.
We've tested the first ROV int he water. We're having some trouble with our sensors, but we hope to have that figured out by the end of the week.

The first ROV:
ROV1
ROV1
0224121844.jpg (163.81 KiB) Viewed 10488 times
2
2
0224121844c.jpg (167.88 KiB) Viewed 10487 times
0224121845b.jpg
0224121845b.jpg (183.07 KiB) Viewed 10487 times
The pipes at the top were the mockups of our ballast tanks. Unfortunately, due to time constraints we won't be able to get working ballasts on our robots.

Here's a video of the first vehicle in the pool:

It can also dive to the bottom of the pool (about 12.5 feet) just fine.


Our goal is to have two robots autonomously locate and move a long pipe in a swimming pool. Unfortuantely, we're rapidly running out of time, so the 'moving the pipe' portion of the task will have to be shown in simulation. However, we think we can get collaborative search figured out by the end of the month. Here's an image we took of the target pipe underwater. The picture was taken using the cheapest network camera we could find.
test_0698.jpg
test_0698.jpg (13.56 KiB) Viewed 10487 times

We're also working on making our second vehicle. At this point the hull is ready to be pressure tested. After that, there's just a little bit of work left with the electronics until the second vehicle is ready to be put in the water (in ROV form, we're still working on sensors).
0322121957a.jpg
0322121957a.jpg (182.46 KiB) Viewed 10487 times
The second robot after paint:
0323121438b.jpg
0323121438b.jpg (206.49 KiB) Viewed 10487 times



Other misc. updates:
We put our wireless router inside the vehicle and ran an antenna extension up to the surface. I was skeptical, but it turns out that this method actually works quite well. As long as the antenna on the surface doesn't go underwater, we can reliably communicate with the vehicle. We've only tested this with a 13' antenna extension, so I can't promise that it will work for deep diving vehicles, but it would be a good solution for shallow water vehicles that are looking to get rid of the tether.

Like I mentioned in another thread, we tried out a cheap waterproof ultrasonic sensor from ebay (http://www.ebay.com/itm/Waterproof-Ultr ... 0721354139) and it works fairly well in the downward direction once the circuit board is waterproofed(we just encased ours in epoxy resin). It gives accurate measurements up to at least 12', and interfaces easily with a microcontroller. We're still working on a longer range forward looking sensor, which we will use for localization within a swimming pool.

We've also successfully sent commands to two different microcontrollers connected to wireless routers on the same network. If anyone is interested in controlling multiple rovs at the same time, this is a relatively painless way to do it. We run a serial to network gateway on each router, that way we can send serial commands to the arduino over a wireless network.

Thats all for now, hopefully I'll have more pictures and videos in the near future.

Cheers,
-Seawolf
Post Reply