Decentralized UAV Swarm Control for Multitarget Tracking using Approximate Dynamic Programming

Md Ali Azam, Shawon Dey, Hans D. Mittelmann, Shankarachary Ragi

Research output: Chapter in Book/Report/Conference proceedingConference contribution

1 Scopus citations

Abstract

We develop a decentralized control method for a UAV swarm for a multitarget tracking application using the theory of decentralized Markov decision processes (Dec-MDPs). This study develops a UAV control strategy to maximize the overall target tracking performance in a decentralized setting. Motivation for this case study comes from the surveillance applications using UAV swarms. Decision-theoretic approaches are very difficult to solve due to high dimensionality and being computationally expensive. We extend an approximate dynamic programming method called nominal belief-state optimization (NBO) to solve the UAV swarm control problem for target tracking application. We also implement a centralized MDP approach as a benchmark to compare the performance of the Dec-MDP approach.

Original languageEnglish (US)
Title of host publication2021 IEEE World AI IoT Congress, AIIoT 2021
EditorsRajashree Paul
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages457-461
Number of pages5
ISBN (Electronic)9781665435680
DOIs
StatePublished - May 10 2021
Externally publishedYes
Event2021 IEEE World AI IoT Congress, AIIoT 2021 - Virtual, Seattle, United States
Duration: May 10 2021May 13 2021

Publication series

Name2021 IEEE World AI IoT Congress, AIIoT 2021

Conference

Conference2021 IEEE World AI IoT Congress, AIIoT 2021
Country/TerritoryUnited States
CityVirtual, Seattle
Period5/10/215/13/21

Keywords

  • ADP
  • decentralized MDP
  • multitarget tracking
  • Swarm intelligence

ASJC Scopus subject areas

  • Artificial Intelligence
  • Computer Networks and Communications
  • Computer Science Applications
  • Health Informatics

Fingerprint

Dive into the research topics of 'Decentralized UAV Swarm Control for Multitarget Tracking using Approximate Dynamic Programming'. Together they form a unique fingerprint.

Cite this