Getting some inspiration…
|March 21, 2011||Posted by Michael under Refining the project||
The AR.Drone is quite a recent device, given that it became available for sale last August (2010). It nonetheless did not prevent lots of developer to eagerly jump on the opportunity of programming this new hi-tech gadget. Soon the web began to offer lots of inspirational ideas about the possibilities provided by the quadricopter.
Considering the project we are about to make and the time it will take, it was deemed necessary to have a comprehensive look at what has already been done, in order :
- to start getting a good idea of what is clearly possible to achieve with the Drone, and what might be more complicated
- to use pre-existing ideas to develop our own
- to make sure that our project has not already been entirely done elsewhere
- to look at other approaches used to solve parts of our problem
- to get feedback and experience from people who have spent more hours working on the device than us
- to, in general terms, know the state of the art regarding the development of applications with such a device
A good start may be to have a look at some official videos provided by Parrot, the French company that build the AR.Drone. Despite to commercial aspect of those videos, it is a good way to have a first feeling of the assumed possibilities of the device. More videos and information can be found on the official webpage.
|Indoor flight demonstration||Outdoor flight demonstration|
Since image analysis appealed to us in the first place, and since it is at the same time one of the main fields explored by developers working on this platform (the drone embeds indeed two cameras), we had to have a look on pattern recognition et tag tracking with a drone.
|Object tracking with a computer running an URBI program.||Vertical camera analyzing objects over a desktop.
Provide a good impression on the resolution of the camera
and its ability to recognize patterns
(in fact, the points of interest of an image).
The ROS.org website that host the Robot Operating System project is bustling with plenty of new inspirational ideas that involve all kinds of robots – it’s basically an open source meta operating system for robots. Some of them are really good at showing what a drone (not necessarily our AR.Drone) can do with image analysis enabled.
|AR.Drone following a 2D tag using ROS.||Using Kinect (a 3D enabled camera) on a quadricopter to control altitude and avoid obstacles.||Autonomous flight of an UAV using Kinect.|
The ROS 3D contest page is also worthy to be read, at least for most of the entries. It is mostly related to interfacing robots (not only UAV) with Kinect, the camera made by Microsoft that enables 3D video-caption; however some examples helped us in defining the topic of our master thesis, that will involve a flying device, robots on the ground and image analysis.
At last, but not least, our last discovery is dronehack, a weblog compiling lots of information, news and programs recently published online about the Parrot AR.Drone. We may want to keep an eye on it during the process of our very own project.