Planning for the Real World:
The promises and challenges of dealing with uncertainty


Drew Bagnell, Carnegie Mellon University, Pittsburgh PA USA
Joelle Pineau, Carnegie Mellon University, Pittsburgh PA USA
Nicholas Roy, Massachusetts Institute of Technology, Cambridge MA USA


As autonomous systems move towards increasingly complex real-world domains, new challenges arise for planning and control. Perhaps the most critical aspect is the problem of integrating uncertainty about state into the control loop. The overwhelming intractability of obtaining optimal solutions to these problems has been a major barrier for the deployment of planning systems that reason about uncertainty. In recent years, however, more scalable approaches have been proposed and some have been applied to practical problems. In light of these novel approaches, the goal of this workshop is to discuss important questions and open issues:
  1. What are currently the best techniques for planning under uncertainty?
  2. What representations of uncertainty are most appropriate for real world problems?
  3. What are the remaining frontiers, in terms of algorithmic challenges, for planning uncertainty?
  4. In which real world problems can we gain traction by applying these techniques that manage uncertainty?
  5. What role does learning have to play with regard to both efficient solution and representation of uncertainty?
Main Page
Call for Posters