Distributed Coordination of Autonomous Swarms with Limited or Absent Communication and Intermittent

Abstract

Unmanned maritime, ground, and aerial systems have proven their worth as useful and cost-effective assets in military missions, including intelligence, surveillance, reconnaissance, and target tracking. Coordinated swarms of autonomous platforms could very efficiently and affordably conduct a range of naval operations, including persistent monitoring and patrol, environmental sensing and inspection, cargo delivery, rescue operations, disruption and tracking of potential adversaries, and ensuring access to maritime domains. Further, such autonomous swarms have the potential to cover a larger area in a shorter time, remove military personnel from danger, reduce human workload, and perform riskier missions than manned systems. Although the technology to develop autonomous multi-agent systems has been progressing, open problems remain in the cooperative control of such systems for successful mission execution. It remains a challenge to deploy autonomous platforms that can perform tasks in realistic, GPS-denied environments, where radio communications are limited or unreliable or when inter-robot communication need to be minimized to conserve power during long-duration missions or to reduce the possibility of detection by hostile entities. Moreover, control frameworks for autonomous swarms should be scalable with the number of agents. To address some of these challenges, this project aims to develop fundamental theory and efficient algorithms for distributed coordination of autonomous swarms with limited or absent communication and realistic assumptions on sensing/perception under field conditions such as occlusion or glare and limited sensing range. In particular, motivated by non-verbal (non-explicit) communication in the form of body language or gestures in human collaborations, we hypothesize that the motion/trajectory taken by a certain agent can be information-bearing and hence, we propose to develop novel intent-expressive/legible motion planning and intent estimation algorithms for improving distributed control and coordination of swarms. The proposed cooperative autonomous swarm technology will enable autonomous naval/maritime teams to provide and interpret “non- verbal” cues or motions of communication-less or communication-limited but cooperative agents. Key innovations of this research include: (1) Hybrid system modeling framework for capturing limited/absent communication and realistic assumptions on sensing/perception under field conditions; (2) Intent-expressive motion planning and limited communication strategies to render agent motions information-bearing for enhanced coordination without explicit communication; (3) Intent estimation/prediction algorithms to infer the possible intents of other agents using intermittent data with limited or no communication; (4) Distributed estimation algorithms for inferring and predicting states including for time intervals when there is no data (either from line of sight or communicated information); (5) Distributed cooperative control and coordination strategies for achieving spatiotemporal task specifications. Correspondingly, distributed coordination strategies can be advanced to leverage intent-expressive motions or “non-verbal” cues for missions with limited or absent communication as well as in the presence of intermittent sensing or perception data. Anticipated research outcomes will contribute to foundational knowledge for autonomy and decision-making in complex multi-agent systems, and for improving the safety, predictability, and reliability of current and future systems, which are crucial to Naval mission success and are aligned with the R&D Framework Priorities of Integrated & Distributed Forces, Sensing & Sense-Making, and Operational Endurance. These autonomy algorithms will enable efficient coordinated operation of naval systems over long durations in uncertain, challenging environments.

Document Details

Document Type
DoD Grant Award
Publication Date
Dec 06, 2022
Source ID
N000142312093

Entities

People

  • Sze Zheng Yong

Organizations

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

Tags

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 - DoD AI Strategy
  • Autonomy
  • Autonomy - Autonomous System Control
  • Autonomy - UAVs
  • Space
  • Space - Spacecraft Maneuvers