This is the first in what will hopefully be a series of self-sustaining autonomous robots that would be able to catalogue and sense their environment. Using an array of solar panels, arduino, and small camera, I was able to have a 'robotic' blimp take pictures and power itself.
My initial intention in creating this was to make something that you could set forth in the world and pretty much forget about, and then after some period of time (hours, days, weeks) this thing would come back to you or you'd search it out and you would be able to see all it's little adventures. Maybe even you could If you built enough of these, you could 'crowd' source any sort of large scale data collection job. For instance if you wanted to catalogue the beds of large lakes or the inside of caves or cloud formations. You could also have it just catalogue wildlife in remote areas. Or take one more step toward singularity.
PROCESS PART I
The process started with my initial idea to make a little robot army. I wanted to the robots to be autonomous and survive without human assistance. Similar to Beam Bots, in fact probably exactly like the same, but their aesthetic revolves more on their function then their form. I guess when I get a little better at the process of making these things form will take a much stronger role.
When I first started out, I wasn't sure what kind of robot I wanted to make. I had just, unsuccessfully, drowned a web camera in the Newtown Creek and was thinking about making a sustainable submarine. The obvious obstacle being how would I go about making it sustainable. I did have some ideas on that though. Placing solar panels on the submarine and letting a battery bank charge. When the batteries were fully charged it would open a chamber that would allow water in and sink the submarine. Once the submarine was submerged it would travel around, explore, and take pictures. Once the battery was a quater full or so it would power itself back to the surface. At the surface it would get rid of it's load water and start charging again, to repeat the cycle. And while I still want to try to make something like this, I think time and my abilities would be serious factors in making this happen.
So I moved on to a different idea. Alex Kozovski gave me the idea of using Klann mechanisms for the design of the robot. Using legos and some animated gifs online I tried to fabricate it. The designs were great because with just one rotational movement you would be able to control multiple legs very elegantly. Unfortunately I was having some trouble, because the dimensions of the legos I was using weren't quite to the specifications of the Klann designs. I was able to get the model working relatively adequately, but there were a fair amount of modifications and iterations I would need to make in order to make it work better, and I wanted to focus my attention on it's function.
Around this time I decided that I was going to pursue making a sustainable blimp. The major advantages of making the blimp were that the mechanics of it were pretty simple (in comparison to the other robots) and I had already made one last semester so I was confident that I would be able to produce one reasonably well.
INSTRUCTIONS - PROCESSS PART II
My idea was pretty simple. I wanted to make a blimp with a camera that was solar powered. If weight and time permitted I wanted to add some motor control and GPS so that it could travel on it's own. I did start working on the that idea and hopefully get something like that in the next iteration of the project.
- 4"x12"x1/4" balsa wood
- Link Sprite Camera
- Micro SD Shield
- Arduino Uno
- 4Gb Micro SD card
- 250' of yarn
- 4'x10'x0.005mm Mylar
- Vinyl, Fabric, and Plastic Adhesive
- Solar Panels delivering at least 5 Volts and 150mAs (I made an array of panels to produce the power I needed)
- Stranded Wire, Electical Tape, Soldering Iron, etc.
The first thing I did was to hook up the arduino with the micro SD shield and Link Sprite Camera to make sure that all these things were working. I went to THIS SITE for the initial code and wiring to get the camera up and running. It was pretty simple and worked exactly as the tutorial had promised. Below are some of test shots I was able to get with camera.
Once I was certain that, that would work I needed to modify the code slightly. One issue I was having was that the arduino program was not able to overwrite files on the Micro SD Card. I didn't really want to overwrite files anyway as I wanted to create catalogue of data. So I needed make the name of each file increment up each time it saved it the card. That wasn't too big of an issue, but I because arduino would be on and off a lot, due to the nature of solar power and the sun setting I needed to have some way of setting the last number of the file on memory so that when the arduino and camera rebooted again they would start taking pictures from where they left off. After talking with Jeff Feddersen he suggested that I use the EEPROM library on the arduino and save the information that way. It turned out beautifully and I just modified my code and got it working. At the bottom of the post is the link to the zip file of the arduino code I used.
Once the code was working, I started building the blimp itself. I started wit the housing for the camera, arduino, and micro SD shield. I made a box out of balsa wood with some cut outs for the camera and mounting the arduino. I used balsa wood because it is extremely light, but fairly sturdy for it's weight. However it isn't the best it terms of withstanding weather conditions and things, so for the next iteration I would probably use some for of plastic or stronger wood.
I did some calculations using the power comsumption of the camera, arduino, and micro SD shield, and figure out that I would need about 5 volts at about 150 mA for this device to work. I took the solar panels that I 'liberated' from the sustainable energy surplus cache in the resident's office and tried to come close to those numbers. Unfortunately I didn't have enough of the solar panels I need to get 5 V and 150mA. I could get one or the the other. While this did frustrate me, because I didn't really need too many more to make it working I decided that I would make the panels suppliment their energy intake into charging a rechargeable 9Volt battery. This worked out relatively well, but added to the weight considerably. Although if I added more solar panels this would have also considerably added to the weight.
Once I got the mechanism working relatively well, I moved on to making the blimp itself. I did some calculation based on figures I found on the internet. Essentially for every cubic feet of helium you can lift 28.2 grams. I was about 392 grams with everything, which comes out to around 14 cubic feet of helium. I had to make a fairly large balloon. The balloon had to be about 3'x5'x1'. Once I drew out the shape I cut it out and glued it with the vinyl,fabric, and plastic loctite. You don't have to specifically use that glue, but you do need to make sure that the adhesive contains Methyl ethyl keton in order to properly bond the mylar. The picture above shows the mylar melting and giving a proper bond.
Once the mechanism and the blimp were complete I tethered the two pieces together with some masonry yarn. I floated the Blimp in the studio I was working at and tested the camera and it's floating capacity. Everything seemed to working fairly well and I decided to send it off on it's first maiden voyage around Greenpoint and Bushwick, Brooklyn.
Below are some photos that the Blimp was able to take. Keep in mind the night shots were taken because the 9Volt battery was already fully charged.