3D Guiding for fixed-wing aircraft

Now, you can have your aircraft tracking a (smooth) 3D parametric trajectory.  This is an extension of the existing GVF in Paparazzi, in fact, now we allow even 2D trajectories with self-intersections.

It is based on our accepted article in IEEE Transactions on Robotics Singularity-free Guiding Vector Field for Robot Navigation

All the instructions to use the algorithm are in the wiki.

Pilot a super rotorcraft!

So far all my results with rotorcraft were in simulation… until now! In this experiment, performed at TU-Delft (thanks Ewoud Smeur!), a team of four rotorcraft behaves as a single unit. The proved stability properties of the whole system allows for applications such as collaborative transportation of objects.

If you want to know the mathematical aspects and why this setup works, then check my thesis.

A wiki entry explaining how you can have the same setup running employing Paparazzi will be available soon!

Multi-UAV capabilities

Video done at ENAC (Toulouse) showing some capabilities of Paparazzi for controlling multiple fixed-wing aircraft.

We have already a wiki entry explaining how to implement the circular formations.

First flights of the hybrid vehicle Cyfoam with Chimera board

The Cyfoam is a hybrid vehicle developed at ENAC Drone Lab. The aircraft is a foam, with a 3D printed fuselage, version of the composite-made Cyclone.

The vehicle is powered by the new autopilot board Chimera! which executes the control algorithms developed by Ewoud from Delft MAV Lab. We are currently aiming at a total autonomous mode, e.g., auto take off and auto landing.

Circle formations of fixed-wing aircraft

We have recently developed and tested a formation control algorithm for fixed-wing aircraft in Paparazzi at ENAC. The position of an arbitrary number of vehicles can be controlled in a circular path. In fact, we are not restricting ourselves to circles but to any closed orbit, such as ellipses, thanks to the guidance vector field that guides the planes.
The algorithm is under more tests, but it should be soon available for the general public. It is quite easy to employ, the user has to declare only the IDs of the planes, the communication topology (neighbors’ relationships) and the desired inter-angles. A detailed explanation will be posted soon in the wiki.
In the following video the planes exchange positions every second. Delays, out-of-date positions (GPS delays), packet losses, etc are expected to be (and actually they are) present. It is quite interesting to remark how robust the algorithm is. According to our calculations the impact of such nasty things are not very important (ofc up to a certain point) for the convergence of the algorithm.