In the heavy-duty physical human-robot collaboration, the robot carries a large load and has a wide operating space and a complex operating environment. If the operator slips, trips or accidentally touches the obstacle, the body will lose balance, and the robot directly executes the operator's intent command will increase the degree of injury, and may even lead to dangerous situations such as workpiece collision and robot damage. Ensuring human operational safety is a prerequisite for achieving efficient collaborative tasks. Hence, it is imperative to approach this from the operator's perspective, ensuring that the robot conducts an operational intent risk assessment based on human motion characteristics before executing human-intended instructions. Subsequently, appropriate response strategies are implemented based on the assessment results to ensure the safety of human operations. Firstly, the human motion state data acquisition system is introduced; Secondly, based on the fuzzy comprehensive evaluation method, a risk assessment model was established by using multi-modal information to analyze three kinds of risk factors in motion; Thirdly, the robot's active response strategies for different risk levels are proposed; Finally, a risk assessment experiment of human-robot collaboration intention is carried out. The experimental results show that the robot using the proposed method can effectively detect abnormal behavior information of humans in the process of human-robot collaboration, and make corresponding decisions according to different levels of risk. It is of great significance for the application of heavy-duty physical human-robot collaboration in complex environments, and provides a way to achieve higher reliability and security in the field of physical human-robot collaboration.

中文翻译：

---

基于模糊综合评价的人机协作动态风险评估及主动应对策略

在重载物理人机协作中，机器人承载负载大、作业空间广阔、作业环境复杂。如果操作者滑倒、绊倒或不小心碰到障碍物，身体就会失去平衡，而机器人直接执行操作者的意图命令会增加伤害程度，甚至可能导致工件碰撞、机器人损坏等危险情况。确保人员操作安全是实现高效协作任务的先决条件。因此，必须从操作者的角度来考虑这一问题，确保机器人在执行人类意图的指令之前，根据人类运动特征进行操作意图风险评估。随后根据评估结果实施适当的应对策略，确保人员操作安全。首先介绍人体运动状态数据采集系统；其次，基于模糊综合评价方法，利用多模态信息分析运动中的三种风险因素，建立风险评估模型；再次，提出了机器人针对不同风险级别的主动应对策略；最后，进行了人机协作意图的风险评估实验。实验结果表明，采用该方法的机器人能够有效检测人机协作过程中人类的异常行为信息，并根据不同的风险级别做出相应的决策。对于复杂环境下重载物理人机协作的应用具有重要意义，为物理人机协作领域实现更高的可靠性和安全性提供了途径。