Robotic manipulation of multiple objects as a POMDP.
Joni PajarinenVille KyrkiPublished in: Artif. Intell. (2017)
Keyphrases
- multiple objects
- manipulation tasks
- reinforcement learning
- multi object
- multiple object tracking
- finite state
- state space
- multiple images
- robotic systems
- dynamical systems
- mobile robot
- partially observable markov decision processes
- partially observable
- model free reinforcement learning
- robot navigation
- particle filter
- real time
- occlusion handling
- data association
- markov decision processes
- belief state
- motion planning
- multiple cameras
- belief space
- viewpoint
- partially observable markov decision process
- complex scenes
- optimal policy
- markov chain
- humanoid robot
- occluded objects
- mean shift
- multi agent