The paper discusses robot-assisted navigation, focusing on shared authority control where the robot and human share control based on the situation. The authors propose a new approach where a human’s behavior is considered correct if it closely resembles what other humans would do in similar situations. This is achieved through machine learning and adaptive control, with the robot varying its intervention based on the human’s actions. The study shows that participants preferred this method over traditional controls. The application of this approach is demonstrated in the context of a robotic rollator assisting people with mild cognitive deficits.
Publication date: 17 Oct 2023
Project Page: https://arxiv.org/abs/2310.11241
Paper: https://arxiv.org/pdf/2310.11241
