Multi-Agent Data Collection in Non-Stationary Environments

#optimisation #multi-agent #robotics.
Share

(Hybrid Seminar)

Please indicate if you are attending phyiscally or virtually when you are registering.

Abstract

Coordinated multi-robot systems are an effective way to harvest data from sensor networks and to implement active perception strategies. However, achieving efficient coordination in a way which guarantees a target QoS while adapting dynamically to changes (in the environment, due to sensors’ mobility, and/or in the value of harvested data) is to date a key open issue. In this work, we propose a novel decentralized Monte Carlo Tree Search algorithm (MCTS) which allows agents to optimize their own actions while achieving some form of coordination, in a changing environment. Its key underlying idea is to balance in an adaptive manner the exploration-exploitation trade-off to deal effectively with abrupt changes caused by the environment and random changes caused by other agents' actions. Critically, outdated and irrelevant samples - an inherent and prevalent feature in all multi-agent MCTS-based algorithms - are filtered out by means of a sliding window mechanism. We show both theoretically and through simulations that our algorithm provides a log-factor (in terms of time steps) smaller regret than state-of-the-art decentralized multi-agent planning methods. We instantiate our approach on the problem of underwater data collection, showing on a set of different models for changes that our approach greatly outperforms the best available algorithms for that setting, both in terms of convergence speed and of global utility.

Note

The seminar is hybrid. Physical room location will be provided after registration prior to seminar day. 

Light drinks and snacks will be provided after the seminar for networking opportunities.

The seminar will also be available via Zoom.

If you are unable to attend physically and wish to participate please register and you will receive an email with the details in the lead up to the event.

Further information

Please contact:

adriel@ieee.org

kyle.millar@ieee.org



  Date and Time

  Location

  Hosts

  Registration



  • Date: 21 Jun 2022
  • Time: 05:30 PM to 06:30 PM
  • All times are (UTC+09:30) Adelaide
  • Add_To_Calendar_icon Add Event to Calendar

Physical room location and virtual Zoom Link to be provided after registration.

  • The University of Adelaide, North Terrace Campus
  • https://adelaide.zoom.us/j/84464256298?pwd=RDRVZ2RNM1Q5aTZOaGVuMTFHeVZadz09
  • ADELAIDE, South Australia
  • Australia 5005
  • Building: Engineering and Maths Building
  • Room Number: EMG06 or Zoom

  • Contact Event Host
  • Starts 09 June 2022 05:30 PM
  • Ends 21 June 2022 05:30 PM
  • All times are (UTC+09:30) Adelaide
  • No Admission Charge
  • Menu: Please select if you are attending physically., Please select if you are attending virtually., Please select if you are unsure.


  Speakers

Nhat Nguyen of The University of Adelaide

Topic:

Multi-Agent Data Collection in Non-Stationary Environments

Coordinated multi-robot systems are an effective way to harvest data from sensor networks and to implement active perception strategies. However, achieving efficient coordination in a way which guarantees a target QoS while adapting dynamically to changes (in the environment, due to sensors’ mobility, and/or in the value of harvested data) is to date a key open issue. In this work, we propose a novel decentralized Monte Carlo Tree Search algorithm (MCTS) which allows agents to optimize their own actions while achieving some form of coordination, in a changing environment. Its key underlying idea is to balance in an adaptive manner the exploration-exploitation trade-off to deal effectively with abrupt changes caused by the environment and random changes caused by other agents' actions. Critically, outdated and irrelevant samples - an inherent and prevalent feature in all multi-agent MCTS-based algorithms - are filtered out by means of a sliding window mechanism. We show both theoretically and through simulations that our algorithm provides a log-factor (in terms of time steps) smaller regret than state-of-the-art decentralized multi-agent planning methods. We instantiate our approach on the problem of underwater data collection, showing on a set of different models for changes that our approach greatly outperforms the best available algorithms for that setting, both in terms of convergence speed and of global utility.

Biography:

About the Speaker

Mr. Nhat Nguyen received his Bachelor of Engineering (first class Hons) in Electrical and Electronic Engineering from the University of Adelaide in 2021. He started his PhD degree at the School of Computer Science, the University of Adelaide in March 2021. His PhD research topic is Hannan Consistent Countermeasures Against Non-Stationary Adversarial Attacks on Autonomous Cyber Operations.