Safe and Stable Neural Network Dynamical Systems for Robot Motion Planning

Abstract

Learning safe and stable robot motions from demonstrations remains a challenge, especially in complex, nonlinear tasks involving dynamic, obstacle-rich environments.