Negotiation-based cooperative planning of local trajectories

Abstract

In this work, a cooperative local trajectory planner based on negotiation theory for human‐robot interaction is developed. It is implemented on a robot, which accompanies patients to examination rooms as part of the HoLLiECares project. For this purpose, an existing human–machine cooperation model for decision-making in one-time conflict cases is applied to a time-repeated negotiation of motion primitives. In negotiation theory, time pressure in the form of deadlines is classically used to achieve agreements. Since deadlines do not naturally exist in all technical applications and their artificial insertion would create an unintuitive system behavior for an involved human, a reciprocal tit-for-tat strategy for the automation is applied in the present work to achieve agreements. This leads to a system behavior that is able to dynamically change between human-in-the-lead behavior or automation-in-the-lead behavior and everything in between depending on the concession of the human and thus on human’s desire. The cooperative negotiation-based local trajectory planner is tested simulatively.