Topic 17 of 17

Safety & evaluation in robotics

Safety and evaluation asks whether a robot policy fails safely and how anyone would know: formal verification, safety-critical control, failure detection, and the honest-measurement problem of small-sample real-robot experiments.

arXiv query: cat:cs.RO AND (abs:"safety-critical" OR abs:"safe control" OR abs:"formal verification" OR abs:"failure detection" OR abs:"control barrier")