Artificial Intelligence

Polyvionics, through its partnership with the LSIIT (UMR 7005, CNRS - University of Strasbourg) and Pr. Pierre Collet, has access to the best world technology in terms of Artificial Intelligence: Computers clusters, GPU computation... and so, to a big field of applications

Classical Optimization
Static optimization used for classical problems optimization

MultiDisciplinary Optimization
Static optimization used for systems conception

Static/dynamic optimization for systems nonlinear State-Space modelling (with starting model architecture, or without)

Dynamic optimization for trajectory optimization and Model Predictive Control

Exemple of a real time trajectory optimization for an aircraft turn


Polyvionics is a specialist of the automation field, and strong partnerships give it access to advanced competences and approachs for navigation, identification and control purposes...

Navigation and State observer
Polyvionics has the capability to perform State Observer (especially through the using of Extended Kalman Filters), and its application to aircrafts navigation.
These competences are currently used for the purposes of the SARINA project, for the European Defence Agency

Polyvionics has an important background in terms of model identification, especially for airplanes and quadrotors.
It has the competences to build nonlinear State-Space models from telemetry files, with classical automation methods, or with new methods coming directly from artificial intelligence.

Polyvionics already implemented the following control aproachs for autonomous systems conductions: Classical State feedback, Optimal control, Model Predictive Control, Adaptative control...

Exemple of an autonomous landing algorithm

Autopilot engineering/design

Polyvionics developed several autopilots in the past (Micav 1.0, Micav 1.1, Micav 2.0 (picture, used by ICS (Skyview OpenPilot)),...)
During these autopilots conception and making, a good knowledge in embedded electronic has been acquired

Embedded program
During the Micav autopilots serie development, an important experience has also been aquired in the filed of computing: Embedded program (sensors interfacing, automation algorithms programing...), communication (Ground/Air network making... ), and ground software through 2 ground stations developed by Polyvionics (UCTS 1.0, UCTS 2.0)

Simulation development

Matlab/Simulink simulation design
These competences are currently used for the purposes of the SARINA project, for the European Defence Agency

Hardware-in-the-loop simulation design
This approach consists in connecting a real autopilot to a computer, to be able to "virtually" fly an airplane simulation. This permits to test all the implemented algorithms, and perform an embedded program debugging...
Polyvionics also developed an Hardware-in-the-loop system (For its former Micav 2.0) which has been used by the CNES (Space French Agency), for the purposes of its Perseus project

UAV platform design

3D model development
The Squirrel MAV has been developed with the CATIA conception program, and several other projects are based on this tool

Model making
The Squirrel MAV is made in glass/carbon fiber, and several other projects are based on this making.

Haut de page



Stéphane Querry, PhD candidate at Polyvionics, presented a part of his research work about aircraft piloting through artificial intelligence, in a French scientific event called “Fête de la science”, through a LSIIT (UMR CNRS 7005 – University of Strasbourg ) stand.

Polyvionics is proud to be involved, inside a consortium led by the Italian company “Ingegneria dei sistemi”, in a contract with the European Defence Agency (EDA) for the feasibility analysis of an innovative navigation system (based on SAR systems) for UAV and cruise missiles.

The “Ponts et Chaussées” engineering school has launched several UAV projects, and has chosen Polyvionics for the avionics part development of their innovative systems.

Artificial Intelligence has a very important role in the Polyvionics research, and a very close partnership with the LSIIT (UMR CNRS 7005 – University of Strasbourg ) has been settled for the use of AI in the aerospace science: Static optimization for systems conception, and dynamic optimization for autonomous systems conduction.