Safe POMDP online planning among dynamic agents via adaptive conformal prediction
Online planning for partially observable Markov decision processes (POMDPs) provides efficient techniques for robot decision-making under uncertainty. However, existing methods fall short of preventing safety violations in dynamic environments. This letter presents a novel safe POMDP online planning...
| Egile Nagusiak: | , , , , |
|---|---|
| Formatua: | Journal article |
| Hizkuntza: | English |
| Argitaratua: |
IEEE
2024
|