Distributed reinforcement learning for flexible UAV swarm control with transfer learning capabilities

Federico Venturini*, Federico Mason, Francesco Pase, Federico Chiariotti, Andrea Zanella, Michele Zorzi, Alberto Testolin

*Corresponding author

Research output: Contribution to book/anthology/report/conference proceedingArticle in proceedingResearchpeer-review

Abstract

Over the past few years, the use of swarms of Unmanned Aerial Vehicles (UAVs) in monitoring and remote area surveillance applications has become economically efficient thanks to the price reduction and the increased capabilities of drones. The drones in the swarm need to cooperatively explore an unknown area, in order to identify and monitor interesting targets, while minimizing their movements. In this work, we propose a distributed Reinforcement Learning (RL) approach that scales to larger swarms without modifications. The proposed framework can easily deal with non-uniform distributions of targets, drawing from past experience to improve its performance. In particular, our experiments show that when agents are trained for a specific scenario, they can adapt to a new one with a minimal amount of additional training. We show that our RL approach achieves favorable performance compared to a computationally intensive look-ahead heuristic.
Original languageEnglish
Title of host publicationProceedings of the 6th ACM Workshop on Micro Aerial Vehicle Networks, Systems, and Applications
Number of pages6
PublisherAssociation for Computing Machinery
Publication dateJul 2020
Article number10
ISBN (Electronic)9781450380102
DOIs
Publication statusPublished - Jul 2020
Externally publishedYes
Event6th ACM Workshop on Micro Aerial Vehicle Networks, Systems, and Applications, co-located with MobiSys 2020 -
Duration: 15 Jun 2020 → …

Conference

Conference6th ACM Workshop on Micro Aerial Vehicle Networks, Systems, and Applications, co-located with MobiSys 2020
Period15/06/2020 → …

Keywords

  • UAV networks
  • distributed deep RL
  • multi-agent RL
  • surveilling

Fingerprint Dive into the research topics of 'Distributed reinforcement learning for flexible UAV swarm control with transfer learning capabilities'. Together they form a unique fingerprint.

Cite this