Header

UZH-Logo

Maintenance Infos

Deep Drone Acrobatics


Kaufmann, Elia; Loquercio, Antonio; Ranftl, Rene; Mueller, Matthias; Koltun, Vladlen; Scaramuzza, Davide (2020). Deep Drone Acrobatics. In: Robotics: Science and Systems, Online, 12 July 2020 - 16 July 2020.

Abstract

Performing acrobatic maneuvers with quadrotorsis extremely challenging. Acrobatic flight requires high thrustand extreme angular accelerations that push the platform to itsphysical limits. Professional drone pilots often measure their levelof mastery by flying such maneuvers in competitions. In thispaper, we propose to learn a sensorimotor policy that enablesan autonomous quadrotor to fly extreme acrobatic maneuverswith only onboard sensing and computation. We train the policyentirely in simulation by leveraging demonstrations from anoptimal controller that has access to privileged information. Weuse appropriate abstractions of the visual input to enable transferto a real quadrotor. We show that the resulting policy can bedirectly deployed in the physical world without any fine-tuningon real data. Our methodology has several favorable properties:it does not require a human expert to provide demonstrations,it cannot harm the physical system during training, and it canbe used to learn maneuvers that are challenging even for thebest human pilots. Our approach enables a physical quadrotorto fly maneuvers such as the Power Loop, the Barrel Roll, andthe Matty Flip, during which it incurs accelerations of up to 3g.

Abstract

Performing acrobatic maneuvers with quadrotorsis extremely challenging. Acrobatic flight requires high thrustand extreme angular accelerations that push the platform to itsphysical limits. Professional drone pilots often measure their levelof mastery by flying such maneuvers in competitions. In thispaper, we propose to learn a sensorimotor policy that enablesan autonomous quadrotor to fly extreme acrobatic maneuverswith only onboard sensing and computation. We train the policyentirely in simulation by leveraging demonstrations from anoptimal controller that has access to privileged information. Weuse appropriate abstractions of the visual input to enable transferto a real quadrotor. We show that the resulting policy can bedirectly deployed in the physical world without any fine-tuningon real data. Our methodology has several favorable properties:it does not require a human expert to provide demonstrations,it cannot harm the physical system during training, and it canbe used to learn maneuvers that are challenging even for thebest human pilots. Our approach enables a physical quadrotorto fly maneuvers such as the Power Loop, the Barrel Roll, andthe Matty Flip, during which it incurs accelerations of up to 3g.

Statistics

Downloads

3 downloads since deposited on 17 Dec 2020
3 downloads since 12 months
Detailed statistics

Additional indexing

Item Type:Conference or Workshop Item (Paper), refereed, original work
Communities & Collections:03 Faculty of Economics > Department of Informatics
Dewey Decimal Classification:000 Computer science, knowledge & systems
Language:English
Event End Date:16 July 2020
Deposited On:17 Dec 2020 08:55
Last Modified:17 Dec 2020 20:30
Publisher:Science and Systems
OA Status:Green
Other Identification Number:merlin-id:20316

Download

Green Open Access

Download PDF  'Deep Drone Acrobatics'.
Preview
Content: Accepted Version
Filetype: PDF
Size: 7MB