Autonomous Space Robotics

Associate Professor T. D. Barfoot
University of Toronto
Institute for Aerospace Studies
4925 Dufferin St., Ontario, Canada M3H 5T6

Phone: +1-416-667-7719
Fax: +1-416-667-7799
Email: barfoot (at) utias.utoronto.ca
Web: asrl.utias.utoronto.ca/

Education

  • Ph.D. – University of Toronto
  • B.A.Sc. – University of Toronto

Awards and Honors

  • Tier II Canada Research Chair in Autonomous Space Robotics

Research Overview

Prof. Tim Barfoot is an Associate Professor and the Principle Investigator of the Autonomous Space Robotics Lab (ASRL) at UTIAS. He began his position in May 2007, after spending four years at MDA Space Missions, where he developed autonomous-vehicle navigation technologies for both space and terrestrial applications. Prof. Barfoot holds a Canada Research Chair (Tier II) in Autonomous Space Robotics, is an Ontario Early Researcher Award holder, and a Professional Engineer (Ontario).

The purpose of Prof. Barfoot’s research program is to enable scientific exploration by creating advanced autonomy for space robotics. Currently, planetary exploration is the primary focus of his work, particularly aspects of estimation and control for planetary rovers. There are currently four research threads at ASRL: (i) localization and mapping, (ii) path planning and path tracking, (iii) novel robotic concepts, and (iv) field testing.

The Localization and Mapping thread examines how to determine where a vehicle is on a planetary surface (either globally or locally) and how to build maps of the environment. These tasks are difficult due to the lack of a Global Positioning System equivalent beyond Earth. For short-range traverses, visual odometry techniques are being developed, which automatically identify and track natural landmarks to infer robot motion using visual sensors such as stereo cameras. For local worksite operations, a set of techniques to build a three-dimensional map using a laser rangefinder is underway. And, for long-range traverses, global localization techniques are being developed including celestial and orbital-image-based navigation.

The Path Planning and Path Tracking thread examines how to find safe passage for a vehicle through outdoor, unstructured, three-dimensional terrain. This can be done by assessing terrain using sensors onboard a robot and then planning to avoid hazards. However, a major challenge in this area is how best to use the limited computational resources available on a rover. A novel concept called a Network of Reusable Paths is under investigation, which allows a robot to explore and establish a network of safe paths, which can be used to revisit locations reliably. This concept builds on the standard visual odometry pipeline and can be thought of as a physical embodiment of a Rapidly-Exploring Random Tree path planner.

The Novel Robotic Concepts thread looks beyond the nominal scenario of having a single wheeled robot carry out planetary exploration. It seeks to develop robotic architectures that simplify the usual approach described in the other research threads. One possibility is to use a team of robots, which could simplify the localization problem by using one another as landmarks. Another is to have a large beach-ball rover blown along by the strong Martian winds, requiring no motors and little in the way of complex sensing or algorithms. Most recently, a tethered cliff-crawling robot concept is under investigation.

Finally, the philosophy of ASRL is that robotics techniques should be proven through realistic field trials. With support from the Canadian Space Agency’s analogue program, Dr. Barfoot carried out preliminary robotic explorations of the Haughton Crater on Devon Island in the High Arctic, in the summers of 2008 and 2009. Additional field tests have occurred in the Sudbury (Ontario) and Mistastin Lake (Labrador) Impact Craters as well as the Mars Emulation Terrain at the CSA in Montreal.