Who wants to be able to swim with your coral?

halmus

Registered Users
M.A.S.C Club Member
ex-officio
#1
I'm in the process of finishing my degree back at UW in Washington while I move my wife out here to Denver. I'll officially make the move here this summer when I'm done. In the mean time, I've been working with a team of guys and have convinced them to take on a marine aquarium related project for our senior project. I thought I would share it here with the community to get some feedback and get to know you all before I can actually make the move and start setting up my next tank and attending meetings. (Here's a link to my first introduction a few months ago: New to Colorado, not new to the hobby)

We're in the Computer Engineering department at University of Washington Tacoma (UWT) and the requirements for our project are a little nebulous. Basically, we are to find a real world problem, and engineer a solution by integrating hardware and software components, including some micro-processor.



Here's the general proposal:

Many hobbyists have cameras they can access remotely (IP cameras, or USB cameras set up through something like Skype). This allows them a fixed viewing angle of their tank. What if they had a system mounted over their tanks that allowed them to move the camera throughout the tank?

Our prototype would be a 2 (or 3) - dimensional axis system similar to a CNC design that would fit below the lighting but could traverse throughout the tank. The possible third axis would allow the camera to change depth of the camera fixture in the water. The camera would also be able to rotate within the acrylic housing and tilt up and down.









A couple notes about the design:
1. This sketchup file shows the initial design which I admit is quite bulky. We would have to optimize the design so that it takes up as little space as possible.
2. Components and drive systems would clearly have to be made out of materials able to withstand the harsh environment of the typical marine aquarium hood with salt-creep and higher humidity.
3. The final design would have to ensure that certain cross-sectional areas of the aquarium be able to be designated "no-fly" zones. By "no-fly" zones, I am referring to regions like overflow boxes, or areas of heavy coral growth that you don't want to run into.
4. The initial design would be a stand-alone unit with a computer monitor and control box (thumb-joysticks) directly tethered to the system. The next step in the design would integrate the control and viewing screen into a computer Graphical User Interface (GUI) which ideally would be accessible remotely through a secure internet connection.

So, the question to the general community is, does this project interest you? If this were a commercially available product, would you consider purchasing a custom fit version for your aquarium? This is, of course, somewhat dependent on the price. We aren't seeking funding. Rather, we are gauging interest in the project to facilitate approval by the committee at UWT that approves or disapproves our project design. To those unfamiliar with the hobby, this sort of project seems like more of a novelty than a commercially viable solution to a problem. Any feedback is much appreciated.
 

halmus

Registered Users
M.A.S.C Club Member
ex-officio
#2
So, as a sort of disclaimer, I am copying the first few postings from another site as this project has been ongoing for a few months now. So, the dialog might seem a little odd, because the timeline is compressed.

We finally placed a couple big orders with some online retailers. 8020 on Ebay is a great source for extruded aluminum and fittings. Servo City is where we got all the gears, bearings, mechanics for the project so far. Circuit Specialists has been our go-to source for stepper motors and electronics components.

Here's our incomplete "parts" table set up in my makeshift shop (what was once my wife's garage).







I had to sort out and store all the little components so I knew what I had to work with.



I ordered 10' of plastic chain. This isn't the most ideal drive system, but we are on a budget and a timeline. This was the quickest and easiest drive system we could put together. CNC's use a type of threaded rod to drive their mechanisms but we don't need the precision that a threaded rod can provide, and the chain will allow us to more easily drive the assemblies back and forth faster. We will be able to set the max speed of travel in the software. We don't want to scare the fish. We also don't want it to take 5 mins to travel across the aquarium.

The plastic chain came in individual links. Imagine putting 10' of this together. After doing some assembly, I'm realizing that I'll have to order another 10'. I can't wait.



This is what we got done in a few hours of work yesterday. The aluminum sections that extend past the horizontal pieces will eventually be cut down. We left them long because we wanted to have plenty of room to work with while we assemble this and work out the bugs. Also, we aren't absolutely set on the final height above the water. That will depend somewhat on the camera housing and how compact we can make it.



To really make this a viable product, it still needs to take up less space. However, we're computer engineers, not mechanical engineers. We just want to get a prototype put together and working. The control systems, electronics, and software are what we're most concerned with. If this was ever really considered as a product, we'd let the big brains in mechanical engineering optimize everything for us.

A side note, after working on this a little, I'm realizing that it's probably not reasonable to have 3-axis on a tank like this. However, I'm still convinced that it is totally do-able to have a single track built into my light rack to allow the camera to move the length of the tank and still be able to see everything. We're just trying to add some complexity to the project for the challenge.

Scaling this up to a bigger tank, the same basic hardware could be used by only extending the aluminum rails. At that point, the rack would start to look less bulky relative to the size of the tank. Either way, our build already is more compact that what we drew up in sketchup. We're using multiple DIY gear systems to provide torque to smaller motors. I'll keep posting pictures as this comes together. The work yesterday gave us a good baseline of what parts we still need and how this will come together as a unit.
 

halmus

Registered Users
M.A.S.C Club Member
ex-officio
#3
This was a quick test to see if the current stepper motor we have on hand will drive this so far. It doesn't have to move much yet, but the test was promising. We were able to put quite a bit of pressure on the linear bearing and not stop the motor.



I'm still hobbling around on crutches, but I have a wooden leg on hand from the last time I broke that foot. It gives me a little freedom in the shop.



We're currently messing around with a Raspberry Pi microcontroller for the video. One of the team had video streaming through it yesterday. It looked good even through an acrylic box I made to submerge in the 40b frag tank I have in the garage.

 

halmus

Registered Users
M.A.S.C Club Member
ex-officio
#5
We've gotten a lot of work done on various elements that will be critical to the project, but they aren't all integrated together yet. It's a balancing act of fitting in time when I can, making it to the laser cutter to cut out some pieces that I overlooked, and waiting for parts in the mail.

We've finally started wiring in the motors and some of the sensors we'll be using to monitor movement.













And, we've got the case put together to hold the power supplies, the microcontroller, and all the supporting circuitry. We still have to machine the fan covers and finish wiring up the actual hand controller once a few parts come in.





The camera housing and all the supporting structure is coming along. We're waiting on a few parts for that also.
 

halmus

Registered Users
M.A.S.C Club Member
ex-officio
#6
Here's an extended version of the laser cutting we did. Some of the parts, like the DIY bearing system, and custom gears needed the type of precision provided by this sort of machine. Sorry about the cheesy music. I had to insert it to cover up some of my commentary that I had intended to edit out.

[video=youtube;tjtX_jOe1LE]http://www.youtube.com/watch?v=tjtX_jOe1LE[/video]​
 

halmus

Registered Users
M.A.S.C Club Member
ex-officio
#7
This is a video of some of the work being done on the camera system now. There's a lot of things to incorporate into a small area here. Movement up and down (z-axis), rotation (partially implemented so far with the DIY bearings and a custom gear that we're hoping to cut out on the laser cutter soon), the tilt of the camera itself, and the microcontroller integration for the camera.

[video=youtube;osOOAT_ni0M]http://www.youtube.com/watch?v=osOOAT_ni0M[/video]
 

halmus

Registered Users
M.A.S.C Club Member
ex-officio
#8
The controller is basically finished now. We are using an Ethernet cable to link it with the main controlling box so that it can be easily removed. We also had to add the glowing logo we adopted for the project.







This is an end view of the aluminum rails we're using for the project. That's where the logo came from.

 

halmus

Registered Users
M.A.S.C Club Member
ex-officio
#9
We're having too much fun with the laser cutter. We designed some fan covers to finish the back of the controller. They're now installed, but need a few more screws.



 

FinsUp

According to my watch, the time is now.
M.A.S.C Club Member
#11
My initial reaction is that the acrylic cylinder isn't needed. GoPro cameras have a waterproof housing, and those would be the ideal cameras to film with. That reduces bulk significantly.
Second reaction is that there's a risk of knocking over coral and rock when mechanically moving something around in the tank, and the majority of my tank would be a no fly zone for that reason.

None of that makes this a less-cool idea, but probably only functional on very large tanks.
 

Bajamike

Reef Shark
M.A.S.C Club Member
#12
I think this is a cool project thanks for sharing. One thing I did notice is why not use a gopro instead of an acrylic tube I think the black is able to wireless.
Its a cool but not sure I would shell out the cash but that's just me. please keep the updates coming I enjoy stuff like this.
 

sethsolomon

Hammerhead Shark
M.A.S.C Club Member
#13
FinsUp!;286206 said:
My initial reaction is that the acrylic cylinder isn't needed. GoPro cameras have a waterproof housing, and those would be the ideal cameras to film with. That reduces bulk significantly.
Second reaction is that there's a risk of knocking over coral and rock when mechanically moving something around in the tank, and the majority of my tank would be a no fly zone for that reason.

None of that makes this a less-cool idea, but probably only functional on very large tanks.

I believe the idea for use with IP cameras to control this from something like an apex or andruio controller.
 

halmus

Registered Users
M.A.S.C Club Member
ex-officio
#14
FinsUp!;286206 said:
My initial reaction is that the acrylic cylinder isn't needed. GoPro cameras have a waterproof housing, and those would be the ideal cameras to film with. That reduces bulk significantly.
Second reaction is that there's a risk of knocking over coral and rock when mechanically moving something around in the tank, and the majority of my tank would be a no fly zone for that reason.

None of that makes this a less-cool idea, but probably only functional on very large tanks.

Thanks for the feedback. I haven't worked with the GoPro cameras. I'll have to look into them. They would be great if they could stream live HD video to the internet. That's our ultimate goal, to embed the live video feed on a web page.

As far as knocking over coral, that's a significant part of the project. We have a couple options, but we aren't planning on letting this fly blind in the tank. I care about coral too much to allow that. There will either be an active sensor system for obstacle avoidance or possibly a system that allows the user to map only certain zones as view-able.

Still, this system clearly isn't practical for all tanks. A more practical option would be to have a single axis traversible in a tank possibly in the front or back. That would limit the space needed for equipment while still providing a lot of viewing options.

We've just added a lot of extra functionality for the challenge.
 

halmus

Registered Users
M.A.S.C Club Member
ex-officio
#15
Bajamike;286208 said:
I think this is a cool project thanks for sharing. One thing I did notice is why not use a gopro instead of an acrylic tube I think the black is able to wireless.
Its a cool but not sure I would shell out the cash but that's just me. please keep the updates coming I enjoy stuff like this.
Thanks for the feedback. I agree that it would be hard to justify spending the money on this. Still, it's a fun project. I intend on integrating some version of this on my future tank. Just a single axis down the center of the tank built into the lighting rack.
 

FinsUp

According to my watch, the time is now.
M.A.S.C Club Member
#16
Halmus - Most of the awesome footage from Animal Planet, Discovery, etc, is filmed with GoPro cams. They're spectacular, and extremely small with multiple mounting options.

Seth- What is this strange language that you speak?
 

halmus

Registered Users
M.A.S.C Club Member
ex-officio
#17
sethsolomon;286209 said:
I believe the idea for use with IP cameras to control this from something like an apex or andruio controller.
Thanks for checking it out. You're right, we are using the Raspberry Pi for the video.

The Go Pro might have been a better option. Ultimately, I don't think it would have saved all that much room. The camera would still need to rotate, tilt 90 deg, and traverse up and down. It would have amounted to about the same footprint.

I originally didn't put much effort into researching IP cameras or other cameras that were ready to run out of the box because I had spent $300 on an IP camera for a former tank that was really disappointing. The R-Pi camera feed is very impressive and cheep. Besides, the R-Pi code is easily modifiable making the video feed integration easier to handle. Still, I'll keep the Go Pro in mind for future projects.
 

MartinsReef

Reef Shark
M.A.S.C Club Member
#18
WOW what a cool read and a great idea. However I think the GoPro med work better but that has been covered. Following this project.
 

Labsalesguy

Angel Fish
M.A.S.C Club Member
#19
I think institutional and zoo aquariums would LOVE to have this item. Imagine the line of kids and adults that would like to try it out! Its like a virtual submarine!
 

MuralReef

Administrator
Staff member
M.A.S.C Club Member
M.A.S.C. B.O.D.
MASC Vice-President
#20
I would love something like this for my classroom tank. This way I would be able to be streaming video 24/7 and have web access for my kids.
 
Top