This paper discusses the importance of safety in autonomous systems, particularly in human-robot interactions. It introduces the Multimodal Safe Set Algorithm (MMSSA) which deals with multimodal uncertainty, a situation where the system is uncertain of the mode it’s in and each mode has its own uncertainty. This is particularly applicable in human-robot interactions where the robot can’t perfectly predict human actions and intentions. The MMSSA is the first energy-function-based safe control method applied to systems with multimodal uncertainty. The paper demonstrates the application of this controller in a simulated human-robot interaction, showing its efficiency and improved safety.
Publication date: 21 Nov 2023
Project Page: Not provided
Paper: https://arxiv.org/pdf/2311.11898
