Robot Teammates Operating in Dynamic, Unstructured Environments (RT-DUNE)
21 May 2018
Robotic systems are beginning to conquer well- to partially-structured environments, such as human-challenging games and self-driving cars operating on well-known streets, but genuinely unstructured and dynamic environments still exceed the capabilities of robotic systems, mainly when robots must also operate with or around humans. This unsolved problem of teams of robots that are capable of adapting to real, unstructured, dynamic environments in real time alongside humans is of high importance to robotics research today, with significant applications for military, search and rescue, and other robot uses in the real world.
To illustrate, imagine a human-robot team operating in an environment that is entirely unknown, such as under a dense tree canopy, at a disaster site, underground, or underwater, where limited or no resources (GPS, comms, power) are available. Compound the difficulty with dynamic events that alter the environment or robotic perception of the environment, such as weather and environmental effects or the actions of other agents.
Successful operation of a robot or group of robots in these most difficult environments requires solutions at the intersection of several of the most challenging areas in modern field robotics: the ability for robots to 1) perceive, reason, and act in dynamic, unstructured environments 2) alongside human teammates and 3) at human-operational speeds.
This workshop will bring together researchers in areas of perception, cognition, planning, control, estimation, machine learning, multi-robot systems, and human-robot interaction to discuss the challenges of robots operating in these contexts, and current and future solutions. Along with this expertise in the solution spaces for this challenge area nexus, this workshop will also feature experts in defining and understanding the real-world problem space from government and industry groups. This combination will be influential in shaping an active conversation and understanding of the obstacles and solutions to building successful robot teammates.
Nicholas Roy, Massachusetts Institute of Technology
Henrik Christensen, University of California, San Diego
Dieter Fox, University of Washington
M. Ani Hsieh, University of Pennsylvania
Stuart Young, U.S. Army Research Laboratory
Simon Ng, Australia Defence Science and Technology Group
Dirk Schulz, Fraunhofer FKIE
Ethan Stump, U.S. Army Research Laboratory
Carlos Nieto-Granda, University of California, San Diego
Christopher Reardon, U.S. Army Research Laboratory
Questions and correspondence should be directed to christopher.m.reardon3.civ at mail.mil.