Robot Task Planning
Overview
Robot task planning studies how a robot can turn an instruction or mission into ordered actions that fit its abilities, surroundings, and physical constraints. Approaches range from symbolic planners such as PDDL, behavior trees, and temporal logic to language- and vision-language models, learned skills, and world models; many combine these with motion planning or formal verification. Hierarchical plans, scene representations, and feedback-driven replanning are used across household tasks, navigation, manipulation, and assembly, with benchmarks increasingly testing long task sequences and changing environments. A central challenge is making flexible plans reliably executable: systems must ground language in physical state, recover from errors, and satisfy safety constraints without excessive computation or human intervention.
Latest papers
242 papers45 papers in the last four weeks, up 55% on the four weeks before. 0.3% of all new papers.