ICRA 2014 Workshop on Long Term Autonomy


Achieving long term autonomy in real world situations is an area of focus for many research labs and industries in the robotics community. The Long Term Autonomy workshop has run at ICRA since 2011, each year bringing together members of academia and industry working on large-scale problems in perception and navigation in a range of challenging environments on land, underwater and in space. 

This year the workshop will focus on the application of Life-Long learning techniques to the problem of long-term autonomy—investigating how addressing robotic tasks such as navigation, planning and manipulation as a form of life-long learning problem can bridge the gap between today’s systems and the robust autonomous systems envisaged of the future.

Call for Contributions

We invite submissions of two-page extended abstracts (three pages maximum) describing new and ongoing work, or proposing future research directions, in areas related to life-long learning for long term autonomy. The authors of accepted abstracts will have the opportunity to give a brief (15 minute) talk at the workshop, followed by open discussion among participants. Optionally and additionally, attendees are invited to submit videos of existing and future autonomous systems - videos will be shown during the lunch period.

To encourage debate and the exchange of ideas, each abstract submission should include a short section listing two questions that the authors would like to discuss after their talk. Submit by emailing your extended abstract, in ICRA format, as a PDF file to:

ltaworkshop2014@gmail.com

Submissions will be reviewed by the organizing committee. Specific instructions for video contributions will be posted on the workshop website shortly.

Important Dates

  • Submission deadline: April 1, 2014
  • Notification of acceptance: April 14, 2014
  • Workshop: June 1st, 2014, 9 AM - 5 PM

Workshop Location and Scheduling

The LTA workshop is a Full-Day workshop that will take place on Sunday, June 1st 2014.

Intended Audience

The workshop is targeted at academic researchers and industry professionals whose interests lie in enabling long-term service and field robotics. It will include presentations from representatives of both industry and academia with the aim being to understand the current limitations of robot autonomy in the field, to provide a snapshot of the state-of-the-art in academic research and to give shape to future research directions.

Workshop Format

The workshop format will include a mix of presentations from invited speakers and those solicited via a call for contributions.  Substantial time during the day will be allocated for discussion, this, together with a panel session will encourage audience participation. We hope that by allowing plenty of opportunity for discussion and audience involvement the workshop will provide the genesis for collaborations which work towards the goal of developing systems that employ life long learning approaches as a step towards long term autonomy.

Topics

 Topics of interest include, but are not limited to:

  •   Lifelong learning for navigation and mapping
  •   Introspection for failure and novelty detection
  •   Persistent state estimation in dynamic environments
  •   Deep learning, object discovery and reinforcement learning applications
  •   Reasoning about environmental change
  •   Long-term planning and exploration strategies in changing environments
  •   Use of semantic information to improve performance
  •   Leveraging the cloud for life-long learning
Organizers
  • Dr Liz Murphy, George Washington University (liz_murphy(at)gwu.edu)
  • Dr Jonathan Kelly, University of Toronto (jkelly(at)utias.utoronto.ca)
  • Dr Paul Furgale, ETH Zurich (paul.furgale(at)mavt.ethz.ch)
  • Prof Gabe Sibley, George Washington University (gsibley(at)gwu.edu)
  • Prof Tim Barfoot, University of Toronto (tim.barfoot(at)utoronto.ca)