Robot Planning in the Real World: Research Challenges and Opportunities

bouncerarcheryΤεχνίτη Νοημοσύνη και Ρομποτική

14 Νοε 2013 (πριν από 3 χρόνια και 11 μήνες)

73 εμφανίσεις

Robot Planning in the Real World:
Research Challenges and
Opportunities

Satyandra K. Gupta


Program Director

Robust Intelligence

National Robotics Initiative

Division
of Information and Intelligent
Systems
National Science Foundation

Relevant NSF Programs


National Robotics Initiative


Robust Intelligence


Opportunity #1: Integrating Planning at
Multiple Levels of Abstractions


Mission Planning


Task Planning


Path Planning


Coverage Planning


Trajectory Planning


Gait Planning


Grasp Planning


….



Opportunity #2: Perception and
Planning Integration


Perception system performance varies over the workspace


Planning has to consider capabilities of the perception system
and exploit it

Opportunity #3: Learning and
Planning Integration


Planning need to include learning and improve performance


Learning from demonstrations


Reusable plan fragments


Opportunity #4: Human Intention
Aware Planning


Robots interact with other humans in the environment


Planning needs to consider intentions of humans in the
environments


Workshop Goals


Develop a shared understanding of what are the open problems
in robotics planning area


Develop a strategy to make progress


Anticipated Outcome


Publicly accessible workshop report that can be cited by the
community


Target deadline: November 30
th