21-000000236_Reactive Swarm Control for Dynamic Environments

Abstract

Approved for public releaseReactive Swarm Control for Dynamic EnvironmentsPI: Dr. Matthew T. Hale, University of FloridaModern uses of autonomy have created a need for operations in dynamic and unknown environments. For example, a swarm of unmanned undersea vehicles (UUVs) may operate in unmapped territory, possibly including hazards such as the presence of adversaries and environmental obstacles. Mathematically, these aspects of swarm coordination mean that a swarm may never operate at steady-state. Instead, ever-changing conditions require swarms to use control strategies that are reactive, in that they must be able to respond to obstacles, adversaries, and other hazards as they emerge, all while still completing their objectives. Curated laboratory settings are able to solve these problems with real-time control techniques that leverage offboard resources, such as offboard computers and wall-mounted motion capture systems, to provide fast-paced, optimal decision-making. Such infrastructure is of course not available outside the laboratory, and continued reliance upon it will exacerbate a growing gap between what is possible in laboratory environments and what is possible in the real world.In this project, we will develop new reactive, autonomous capabilities for swarms using only onboard capabilities and resources. In particular, only onboard computing and offboard sensing will be used. To do so, we will develop techniques to embed optimization algorithms in feedback loops for swarms, incorporating the flow of information from conventional feedback signals, computations onboard each agent, and communications among agents. Overarching questions to be answered are: 1.Under what conditions on dynamics, computations, and communications is this configuration stable?2. What event-triggering mechanisms should be used to compute new inputs and/or communicate information to other agents?3. To what extent can decision-making of this kind provide some form of optimality guarantees without being able to reliably predict future conditions? All theoretical developments will be extensively tested on teams of robots both indoors and outdoors.

Document Details

Document Type
DoD Grant Award
Publication Date
Jun 09, 2021
Source ID
N000142112495

Entities

People

  • Matthew Hale

Organizations

  • Office of Naval Research
  • United States Navy
  • University of Florida

Tags

Fields of Study

  • Computer science

Readers

  • Agent-Based Social Robotics and Mobile-Assisted Learning in Virtual Environments.
  • Distributed Systems and Data Platform Development
  • Unmanned Aerial System (UAS) Autonomous Capabilities and Mission Reconnaissance.

Technology Areas

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