Human–robot collaboration (HRC) has attracted strong interests from researchers and engineers for improved operational flexibility and efficiency towards mass personalization. Nevertheless, existing HRC development mainly undertakes either human-centered or robot-centered manner reactively, where operations are conducted by following the pre-defined instructions, thus far from an efficient integration of robotic automation and human cognitions. The prevailing research on human-level information processing of cognitive computing, the industrial IoT, and robot learning creates the possibility of bridging the gap of knowledge distilling and information sharing between onsite operators, robots and other manufacturing systems. Hence, a foreseeable informatics-based cognitive manufacturing paradigm, Proactive HRC, is introduced as an advanced form of Symbiotic HRC with high-level cognitive teamwork skills to be achieved stepwise, including: (1) inter-collaboration cognition, establishing bi-directional empathy in the execution loop based on a holistic understanding of humans and robots’ situations; (2) spatio-temporal cooperation prediction, estimating human–robot–object interaction of hierarchical sub-tasks/activities over time for the proactive planning; and (3) self-organizing teamwork, converging knowledge of distributed HRC systems for self-organization learning and task allocation. Except for the description of their technical cores, the main challenges and potential opportunities are further discussed to enable the readiness towards Proactive HRC.

人机协作 (HRC) 已经吸引了研究人员和工程师的浓厚兴趣，以提高操作灵活性和效率以实现大规模个性化。然而，现有的HRC开发主要采取以人为中心或以机器人为中心的反应方式，按照预先定义的指令进行操作，远未达到机器人自动化与人类认知的有效融合。认知计算、工业物联网和机器人学习的人类级信息处理的流行研究为弥合现场操作员、机器人和其他制造系统之间的知识提炼和信息共享差距创造了可能。因此，一个可预见的基于信息学的认知制造范式，Proactive HRC，交互协作认知，基于对人类和机器人情况的整体理解，在执行循环中建立双向同理心；(2)时空合作预测，估计分层子任务/活动随时间推移的人-机-对象交互，以进行主动规划；(3)自组织团队合作，融合分布式 HRC 系统知识用于自组织学习和任务分配。除了对其技术核心的描述外，还进一步讨论了主要挑战和潜在机遇，以便为主动式 HRC 做好准备。