“Follower” to “Collaborator”: A robot proactive collaborative controller based on human multimodal information for 3D handling/assembly scenarios
Hubo Chu,Tie Zhang,Yanbiao Zou,Hanlei Sun
DOI: https://doi.org/10.1016/j.jmsy.2024.02.008
IF: 12.1
2024-04-01
Journal of Manufacturing Systems
Abstract:At present, human-robot collaboration systems usually focus on robot followership and only achieve one-way system feedback. To make robots proactively collaborate with humans in unstructured and unknown scenarios and achieve the transition from "Follower" to "Collaborator", a robot proactive collaborative controller based on human multimodal information is proposed, which integrates a motion planning controller (MPC) for predicting human arm motion and a reaction controller (RC) for learning human perception and variable impedance behavior. In MPC, a fine-grained and lightweight motion prediction model based on human joint angle information is proposed to help robots predict human arm motion and achieve proactive collaboration. In RC, a refined perception interface for human collaboration states based on human surface electromyography signals and an online learning strategy for human variable impedance behavior are proposed to help robots deal with non-synchronous motion and improve collaborative comfort and stability. Furthermore, by imitating human learning mechanisms, an exploratory iterative learning algorithm is proposed to optimize the performance of the learning strategy. Offline and online experimental results validate the effectiveness of the proposed controller. The human-robot collaborative handling and assembly experiments show that compared to other advanced methods, the proposed controller achieves higher collaboration efficiency and comfort.
engineering, manufacturing, industrial,operations research & management science