Human-Robot Co-Training: A Novel Paradigm for Symbiotic Human-Robot Collaboration

Abstract

This project aims at providing fundamental contributions to the topic of training synergistic humanrobot teams that progressively g,row in capability over time. Similarly to the symbiotic relationship between a multipurpose military working dog and its handler, we, will develop a robot agent able to adapt to and monitor its human partner, provide personalized intervention during and after train,ing, and ultimately leverage the superior cognitive and inductive capabilities of the human to maximize unit performance and resilie,nce. Effective teamwork requires cooperation among participants, explicit or implicit coordination of activities and sub-tasks, and,collaboration. However, these terms are too abstract to be actionable, and there currently exist a gap between high-level research i,n cognitive systems engineering for human-machine teamwork and technological interventions/applications. Building on our own experti,se in bridging this gap, we propose to incorporate notions of task allocation and teammate role assignment in a shared autonomy fram,ework.Project Objectives and OutcomesT1 Task Decomposition and Capacity Assessment. Development of task representations that captu,re skill and capacity requirements for dynamic task allocation and assessment, to allow both the human and the robot to perform unde,r different roles and take on a spectrum of responsibilities.T2 Fluid Coordination and HumanRobot Interdependence. Development of, a robot teammate that observes and predicts human behavior to direct human action toward opportunistic collaborations for the benef,it of the team.T3 Human-Robot Co-Training in Shared Autonomy. Extension to a flexible shared autonomy framework in which the human, and the robot can assume (and negotiate) different roles, perform different tasks, and evolve in symbiosis over time. We envision a, dynamic framework that optimizes individual contributions to maximal unit performance.Impact on DoD capabilitiesHumanRobot Collabo,ration has clear potential to streamline operational readiness, preserve the strategic depth of the Navy, and disrupt the battlefiel,d of the future. This project will advance knowledge on the problem of fluid synchronization in humanrobot teaming, and will make p,rogress toward a future where robot collaborators will be used as a strategic advantage against adversaries. We envision three areas, of impact on the Navy and DoD at large: i) logistics, inspection and maintenance of vehicles and buildings, and inventory of wareho,uses and equipment/tools; ii) human-autonomy co-training as a paradigm for next generation training programs in the Navy; iii) the b,attlefield of the future, where sailors and marines are empowered to take advantage of autonomous technologies in novel, creative, a,nd unprecedented ways.Approved for Public Release

Document Details

Document Type
DoD Grant Award
Publication Date
Jul 13, 2022
Source ID
N000142212482

Entities

People

  • Alessandro Roncone

Organizations

  • Office of Naval Research
  • Regents of the University of Colorado
  • United States Navy

Tags

Fields of Study

  • Computer science

Readers

  • Agent-Based Social Robotics and Mobile-Assisted Learning in Virtual Environments.
  • Systems Analysis and Design
  • Team-Based Human-Centered Cognitive Task Decision Making and Information Performance.

Technology Areas

  • AI & ML
  • AI & ML - DoD AI Strategy
  • Autonomy
  • Autonomy - Human-Robot Interaction