Cardboard Hovercraft Robot

I've decided to try out cardboard as an alternative to foam board. The largest incentive for doing this is that I'm now able to use the Laser Cutter to prototype my designs. There's a mantra that I like to abide by when working on projects: fail often and fail fast.

While I like the idea of using foam board, I had to manually cut each layer, which was time consuming and probably unhealthy. On the flip side, cardboard is really _really_ cheap, to the point where Techshop has a whole bunch of the stuff just laying around as scrap. Jackpot!

Kidding aside, I'm still not sure how I want things to fit and be arranged, and am even not sure as to what configuration I'm going for (skirt/no skirt? single/dual hovering EDFs? EDF on a Servo, or Dual EDFs? Control EDF speeds for thrust, or use rudders? etc...). I need to try out a whole bunch of ideas quickly and cheaply (I'm cheap) and cardboard fits the bill.

A main problem that I'll have to address will be cardboard's perforation. If I want to make a chamber (which is very much on the todo list), I'll need a way to seal up the perforations, which is something I didn't necessarily need to worry about with foam board.

Overall, I think the tradeoffs will be worth it, but only time will tell.

The very first proof of concept worked, and then some! I managed to cut some foam board with a makeshift foam cutter (Add one part ATX Power Supply, one part Nichrome Wire, one part wood, and a sprinkle of various metal hardware), and fit two 40mm EDFs onto the part. The LiPo battery and HTC One S were put on there to make sure it would actually list most of the weight.

This test was with three sheets of 1/2" Foam Board, without any type of skirt, and I didn't even need to hit full throttle to get the board to hover. This is very promising since I think performance will only improve once I actually direct the air with some sort of chamber. There was a lot of air leaking through the sides of the craft, which I hope to correct with some sort of chaimber.

I wanted to take this time to explain why I'm using a Smartphone (HTC One S) instead of something else. As I stated in an earlier log, one of the project goals was to re-familiarize myself with Android, but first a bit of backstory.

A few years ago (pre-Gingerbread), I got into Android programming for a job, and one of the things that intrigued me was trying to use my smartphone as a Robot Controller. It has all of the sensors, it has a screen to interact with and get debug information, and it even has a battery. But when I tried using it as such, I found it incredibly difficult, and found that the sensors were sub-par for what I wanted. Fast forward to now, and things have gotten much better. I'd like to flex my Android muscle so that I may use it later when needed.

On a related note, I've been using Linux for many years now, to the point where I put it and use it on as many things as possible. Normally I'd love to use Linux for this project, but there's a bit of a problem. Through a good amount of experience, I've found that all electronics have their own quirks and issues, Linux included. There's a very unique issue that I have to address with this Hovercraft setup in that if I want Linux, it needs to be small and light. There are a few options avaiable such as the Beagle Bone Black and Raspberry PI, but frankly, I would much prefer to use something that was designed to be used by typical consumers instead of hobbyists. To me, the difference is that by using a Smartphone, I only have to worry about the App, as all of the hardware and OS have been through a whole bunch of testing, whereas with a Raspberry PI and BBB, I'd have to add my own hardware, and run a variant of Linux I'm unfamiliar with, AND THEN run my application on top of it.

The goal is to eliminate as many variables from this build as possible, and from a hardware perspective, it's hard to trump Smartphone hardware. The tradeoff is that I likely won't get nearly as much performance out of it as compared to something else, but I've been burned in the past with silly hardware mistakes that I would prefer to avoid if possible.

I have several ideas for attacking the challenge of circumnavigating the Sparkfun AVC course in under 30 seconds (aim for the stars and you'll find planets). From a high level perspective, I'll be separating the Smartphone and MCU functions as follows:

Smartphone:

- Game Logic

- Debug Info

- Computer Vision

- Start/Stop button

Tiva Launchpad (MCU):

- Command Parser

- IMU/Stabilization
- EDF Controller
- Power Monitor

As for the hardware, I'm going to aim for using FoamBoard + Electric Ducted Fans (EDFs) to generate the thrust and hovering. The smartphone in and of itself will unfortunately add a whole bunch of weight to the hovercraft, along with the LiPo battery. I've never used an EDF, let alone a Brushless Motor, so I'm not sure what to expect.

As for the skirt, I've seen several designs online, but very few resources on making your own. First, I'm going to check if I even need one, and if it comes to pass that I need one, I'll deal with that bridge once it's in front of me.

For the past few months, I've been working on a Robotics BoosterPack (Tiva Launchpad equivalent of a shield) aimed to make building robots much easier. Well, I've finally reached a point in which I want to make sure that what I'm making actually makes sense; I want to put the project through its paces.

However, I'd also like to use this as a useful tangent to learn more skills and broaden my experience with things I already know. This is why I'm aiming to make an Autonomous Hovercraft and enter it into Sparkfun 's Autonomous Vehicle Competition.

I have several goals with this project that I'd like to hit:

- Put Robotics BoosterPack through its paces

- Re-familiarize myself with Android programming

- Make a Hovercraft

- Learn how to use and debug Bluetooth, GPS, IMU, OpenCV

- Put everything together and make it all autonomous!

I have no idea whether I'll be able to hit everything on the list, but there's no sense in not trying!