Decentralized Planning for Adaptive Search in Heterogeneous Multi-Agent Systems with Tightly Coupled Tasks

Abstract

This project aims to address problems in coordination planning for outdoor multi-robot systems in which mission objectives and agent s actions are subject to complex constraints, and communications between the agents are severely constrained. The project s aim is motivated directly by the dominant challenges that arise in real-world applications of multi-robot systems, where robots are constra ined by their onboard resources, by their ability to sense, maneuver and communicate, and by the sequence in which different, specia lized vehicles must execute tasks. This class of problems is particularly relevant to teams of underwater robots, where the range, b andwidth and reliability of communication systems are several orders of magnitude less than in terrestrial or above-water domains. The project expects to significantly advance the theory of multi-robot cooperative behavior towards increased capability of practica l systems in the short term and will develop novel decentralized planning and control algorithms for autonomous platforms with the e nd-goal of deploying software on-board a multi-agent system in the field. We will integrate a rich form of temporal logic called sig nal temporal logic (STL) with the decentralized Monte Carlo tree search (Dec-MCTS) framework to tightly couple low-level tasks with high-level mission objectives in a fully decentralized manner. The project s outputs will support a variety of applications includin g mine countermeasures, rapid environmental assessment, situational awareness, and related applications that involve finding a safe passage for a high-value supported by a team of low-value assets through areas where it is exposed to hazards or threats. Our expect ed results will offer increased operational agility through the deployment of underwater multi-robot systems, and will achieve schol arly impact through technical advancements that address some of the most difficult and fundamental questions in the discipline.

Document Details

Document Type
DoD Grant Award
Publication Date
Aug 20, 2021
Source ID
N629092112031

Entities

People

  • Robert Fitch

Organizations

  • Office of Naval Research
  • United States Navy
  • University of Technology Sydney

Tags

Fields of Study

  • Computer science

Readers

  • Agent-Based Social Robotics and Mobile-Assisted Learning in Virtual Environments.
  • Distributed Systems and Data Platform Development
  • Systems Analysis and Design

Technology Areas

  • AI & ML
  • AI & ML - Autonomous Systems
  • AI & ML - DoD AI Strategy
  • AI & ML - Machine Learning Algorithms
  • Autonomy
  • Autonomy - Autonomous System Control
  • Autonomy - UAVs