Improved control UAV’s (unmanned aerial vehicles) for faster, more accurate and robust maneuvering
Projectpartners privé: EUKA, DronePort
Projectpartners kennisinstellingen: KU Leuven, UGent, von Karman Instituut
Projectduur: 4 jaar
Status: In aanvraag
Valorisatie van dit project in samenwerking met EUKA en DronePort:
De valorisatieketen voor UAV-applicaties is een proces in drie stappen. Waarde wordt gegenereerd bij grotere bedrijven en openbare diensten, door hun bestaande bedrijfsprocessen te verstoren. Tal van kleinere bedrijven, zowel bestaande als in de opstartfase, zullen deze ‘verstorende’ processen aanbieden op basis van UAV-hardware die de projectresultaten vervatten. Omdat dit proces steunt op kleinere bedrijven kan dit een snelle, flexibele en robuuste weg naar valorisatie zijn. Dit project versnelt dit proces en begeleidt het om succes te waarborgen.
SBO project proposal for the Research Foundation – Flanders (FWO) – April 2017
Goele Pipeleers, Peter Slaets, Jan Swevers, Maarten Vanierschot, Mark Versteyhe (KU Leuven)
Marc Vantorre (UGent), Jeroen van Beeck (von Karman Institute)
Contact information: email@example.com 0472/ 19 46 98
The main objective of this proposal is to improve the position control of unmanned aerial vehicles (UAV’s). The technology improves position control (accuracy, dynamics, robustness) in proximity to hard solid spatial objects (walls, buildings) and under wind load.
We see a strong need for improved levels of (position) control. At present, a (remote) pilot continuously corrects the intended path of the UAV subjected to disturbances like wind and proximity to large solid objects (such as buildings). Large solid objects like walls cause a push back on the UAV due to aerodynamic interactions with that object. The overall objectives of this project is to assist the pilot boosting the maneuverability performance of UAV’s by improving the controller based on models that are able to predict complex aerodynamic interactions.
Current autopilots use classic linear control techniques for navigation, guidance and control. However, in close proximity to objects or under wind, these aerodynamic interactions become important and traditional autopilot systems fail.
Figure 1 A UAV disturbed by the building as it approaches the building (streamlines of air are compressed and result in a net force on the UAV) and by the flow field (wind) around building. The intended controller can predict these interactions and knows how to pro-actively compensate for: the result is solid position control around objects.
The control architecture of choice is a model predictive controller. An aerodynamic model of a UAV interacting with its fluidic environment is used to predict the result of potential control actions like revving up on of the screws. Being able to predict the result of control actions (over a limited time horizon) enables then the elicitation of best possible set of control actions to apply to the UAV.
New application areas for UAV’s
The project lifts roadblocks on the roadmap toward fully autonomous flight under disturbances.
On a shorter term the project results will allow UAV’s to be operated more precise and with less effort from the pilot. It allows a safer, more robust (against wind), and less risky (against crash or collision) operation, and an easier landing.
The project furthermore enables a number of new indoor applications in which UAV’s need to fly closer to flat surfaces (ceiling, floor, walls), such as close-range inspection, object manipulation with grippers, spray painting objects, cleaning off dust, …
The project also enables new outdoor applications which are currently impeded by wind disturbances that the pilot cannot remotely corrected for. Applications include closer inspection of buildings/pylons, cleaning of window glasses, manipulation by grippers for maintenance, spray painting objects, cleaning or blowing off debris from buildings or structures. The project will also allow operating the UAV at higher wind speeds than is currently possible.