From Semantics to Execution: Integrating Action Planning With Reinforcement Learning for Robotic Causal Problem-Solving.
Manfred EppePhuong D. H. NguyenStefan WermterPublished in: Frontiers Robotics AI (2019)
Keyphrases
- reinforcement learning
- action selection
- plan execution
- durative actions
- plan generation
- partially observable domains
- reward shaping
- planning process
- control knowledge
- state space
- causal knowledge
- control flow
- action descriptions
- function approximation
- macro actions
- bayesian networks
- action theories
- action space
- partially observable
- derived predicates
- solving problems
- dynamic environments
- condition action rules
- real robot
- initial state
- causal graph
- state action
- temporal difference
- mobile robot
- deterministic domains
- indirect effects
- markov decision problems
- planning problems
- reinforcement learning algorithms
- planning domains
- robot control
- machine learning
- motion planning
- ai planning
- reactive planning
- case based reasoning
- logic programming
- heuristic search
- blocks world
- markov decision processes
- action sequences
- robotic systems
- domain independent
- partially observable markov decision processes
- atomic actions
- concurrent actions
- causal relationships
- markov decision process
- causal models
- speech acts
- temporal constraints