Context-aware assistance guidance via augmented reality for industrial human-robot collaboration

Abstract

Industrial human-robot collaboration (HRC) endows industrial robots with the ability to cooperate with workers and complete common production goals. In the HRC scenario, the dynamic and unstructured environment makes it difficult for both humans and robots to collaborate quickly, where an intuitive communication channel for assistance guidance is desired. This paper proposes an augmented reality (AR)-based assistance guidance approach with an improved context-aware recommendation algorithm. The defined HRC-specified context information is extracted and analyzed from the real-time perceived data. The context model is then updated and triggers the guidance event to derive multiple options of assistance guidance views. Finally, a novel context-aware recommendation algorithm based on user-collaborative filtering is proposed to obtain the optimal AR view for assistance guidance. The proposed method is validated in a developed HRC assembly case. The result shows that the proposed method performs better than the other three recommendation algorithms in terms of accuracy, recall, and MAP, which promises to improve the efficiency of human-robot collaboration and reduce the learning cost for HRC tasks.