BeliefSF

Robots gain verifiable safety with new belief-space inference method

Researchers have developed a new method to formally certify the safety of autonomous robots interacting with humans. This approach uses conformal prediction to verify belief-space safety filters, which actively reduce u…