Interactive Autonomy theme

Members: Michael Jenkin (theme leader), John Tsotsos, Dana Kulic, Gregory Dudek, Alexis Lussier Desbiens, Joelle Pineau, David Meger, Angela Schoellig

Partners: CrossWing Inc., Defense Research and Development Canada (DRDC), Barrick Gold, Clearpath Robotics, Canadian Space Agency (CSA), US Open Source Robotics Foundation (OSRF), Kinova Robotics, FPInnovations, Element AI.

The Interactive Autonomy theme is concerned with understanding how groups of robots and mixed human/robot teams can work together safely and effectively. The problem of coordination and cooperation amongst autonomous machines is often phrased in terms of the task of multi-agent coordination and decision making for systems composed of multiple, possibly heterogeneous devices while Human-Robot Interaction (HRI) is a well-established research area that requires the integration of sensing, reasoning and control technology with deep knowledge of human factors and robotics. Key research areas in the Interactive Autonomy theme include:

  • Measurement, prediction and analysis of human motion and actions.In order to operate safely and achieve team performance and fluency, the robot team needs to accurately perceive and anticipate the human location, pose and behaviour, in both indoor and outdoor environments. For extended interactions, the robot system also needs to estimate human physiological and cognitive states, such as engagement, attention and situational awareness.
  • Collaborative perception and action specification.The human-robot team needs to be able to establish shared goals and action plans, without requiring extensive operator knowledge about the robot.
  • Coordinated action and leadership.During execution of shared plans, the human-robot teams need to be able to smoothly and intuitively coordinate action, and fluidly communicate and exchange leadership and following roles.
  • Learning.Experiences obtained through interaction as well as human demonstration should be used to improve future team performance. Developing robots that can adapt to, and learn about, human needs and preferences in shared control.