I'm still getting into the swing of this blogging thing. I figure a blog with no readers is allowed to have a sporadic (read: sparse) update schedule.
Anyway, today's topic is my senior project. If you're an engineering student, you know we're all required to do one (or more) practical project before we graduate--the senior project. Although I'm still in my third year, I have senior standing and I've decided to do my first senior project (of two) this semester. I'm working with three other seniors here, and we're going to take a Parrot AR.Drone, slap a LIDAR on it, and create software to make it able to autonomously explore an indoor environment. We're hoping to get funding for the project from the Air Force Research Laboratory's Student Challenge, which funds senior projects addressing a series of problem statements.
Parrot created the AR.Drone as an Augmented Reality Game toy. It has two cameras onboard, and you fly it using an iP[od|ad|hone]. The iOS app includes games such as shooting other AR.Drones with virtual missiles and such. The cool thing about this is that the computing hardware Parrot uses for this is powerful and modifiable. There's a full embedded Linux machine inside of the AR.Drone, and they left it wide open (didn't even disable telnet). That means that we can start hacking together new code on the drone itself to process our new sensors and maybe tune the control algorithms. Because the Drone communicates over WiFi and Parrot has released an API to control the drone from custom software. Using this API and the Robot Operating System, we're going to write code for mapping (using data from the LIDAR and the AR.Drone's IMU) and autonomous path planning.
Our first challenge is going to be putting all of the hardware together. Parrot designed the AR.Drone very close to the edge of its performance limits. That means it can't take much payload. We want to put a 160g LIDAR on a 420g AR.Drone. Obviously, that's going to take a toll on the hardware. What remains to be seen is how much. People on the internet (always a reliable source) have claimed payloads between 100g and 250g. We're going to lighten what we can wherever we can, but no matter how you slice it, we're walking a fine line between success and destabilizing the drone. We're going to do some stress tests this weekend on a borrowed drone to see how much it can fly with before it loses control, cannot take off, or the motors start overheating.
Sunday, January 23, 2011
Subscribe to:
Post Comments (Atom)
No comments:
Post a Comment