Self-Organized Collective Systems using Implicit Coordination: Closing the perception gap between theory andimplementation for 3D underwater robot swarms

Abstract

Self-organized collectives represent a fascinating approach to autonomy, where seemingly purposeful coherent large-scale behavior emerges from small scale interactions between many individuals -- so much so that the collective appears to be an autonomous entity by itself. Understanding how to harness this method of autonomy has broad implications for engineering human-designed systems to be resilient and scalable. Fish schools have specialized to create massive and impressive levels of collective intelligence in underwater environments,migrating long distances, efficiently searching for resources, and even forming dynamic shapes like bait balls to capture prey. They provide an important existence proof that a high degree of collective autonomy can be achieved underwater, in spite of the inherent sensory and communication limitations. To do so, they rely on implicit coordination: individual fish make decisions based primarily on their visual observation of nearest neighbors and without explicit communication, elegantly bypassing the inherent challenges of underwatercommunication. While many groups are developing methods for above-ground robotic systems, relying on global positioning and wireless communication, the ability to engineer collective behavior underwater through fish-like implicit coordination has been neglected. As a result, theoretical studies of fish-inspired autonomy remain abstract and unvalidated from an engineering perspective.Performer proposes to develop a fundamental understanding of self-organized 3D collectives that use implicit coordination, by utilizing an integrated algorithmic-experimental approach based on a novel 3D underwater robotic platform BlueSwarm. Performer s goal is to close the gap between theory and implementation for bio-inspired implicit coordination collective behaviors, especially with regards to limitations on 3D perception in underwater environments. Our research has two main thrusts: (1) Experimental investigation of four core self-organized behaviors: dispersion, alignment, dynamic formation, and collective capture, and (2) Development of generalized mathematical models for reasoning about perception-limited implicit coordination. Underwater multi-robot systems have many important potential applications: environmental monitoring, inspections of underwater infrastructure, and search-and-rescue operations. Through research, performer aims to advance the state of the art in fully autonomous 3D underwater swarms, as well as advance the theory and algorithmic understanding of realizable collective autonomy with implicit coordination.

Document Details

Document Type
DoD Grant Award
Publication Date
Apr 29, 2020
Source ID
N000142012320

Entities

People

  • Radhika Nagpal

Organizations

  • Office of Naval Research
  • President and Fellows of Harvard College
  • United States Navy

Tags

Fields of Study

  • Computer science

Readers

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

Technology Areas

  • AI & ML
  • AI & ML - Autonomous Systems
  • Autonomy
  • Autonomy - Autonomous System Control