The Hovercraft project consists in developing a robot with a really particular mission: build a map of a changing ground area like passing from hard rock to mud in indoor environment.
In our times, it is not necessary to have read Robots from Isaac Azimov to understand that those artificially crafted creatures will be part of our lives in a not so far future. The only remaining question is: “how?”. According to Azimov, each robot should have a determined mission. This domain is currently having a lot of attention from scientists. We wanted to participate to this great technological adventure by developing a robot with a really particular mission : build a map of a changing ground area like passing from hard rock to mud in indoor environment. Everything should stay on budget of only 300 euros.
Hovercrafts are able to fly on different types of terrain. This is why we choose to build our robot with hovercraft possibilities. We worked on a module created by a previous team and added a better system to it for the designed mission. The hovercraft has a turbine to create an air-cushion under the robot and four motors placed in opposed directions to control robot's displacement.
Our system is composed of two sub-systems: the robot itself and a control interface running on a separate computer. The control interface show the user every robot's variable and the mapping in construction. Both systems communicate through Wifi.
The mapping is done with the original association of a camera and a laser pointer : create an offset between the camera and the laser pointer gives us the possibility to calculate the distance of the object pointed by the laser. The repetition of this measurement with a rotation of the laser pointer and the camera allows us to build a map of the area.
To build a complete mapping, the module must move and know its position in space and time. An hovercraft has no contact with ground so we needed to find a solution in agreement with this constraint. That is why we use a computer mouse sensor, also called optical flow sensor, strategically combined with gyroscopes to determine module position in its displacement plan.
To stay in harmony with our budget constraint, the robot main computation run on the famous micro-pc card Raspberry Pi. For real-time constraints brought by usage of different sensors and motors, we also use the Arduino card. Both cards communicates by a simple USB cable. Moreover, lots of pieces used are from recycled materials.
This project, well conducted in collaboration with the ONERA (the french aerospace lab) shows that it is possible, with a low budget and resources available to everyone, to create a simple autonomous exploration robot.