多源不确定条件下移乘护理机器人多椭球传播分析方法与可靠性研究

A safe and reliable transfer nursing robot plays an important role and has a bright prospect in alleviating the shortage of nursing staff, reducing the burden of children, and improving the quality of life for the elderly people. In the process of propagation and reliability analysis of transfer nursing robot under multi-source uncertainties, there are several bottleneck issues, such as the difficulty of modeling and the poor computational efficiency, due to the lack of nursing experiment data, variables with diverse distribution, complex failure modes, and double-layer nested calculation. This project aims to develop a targeted multi-source uncertainty propagation analysis method and solving strategy on reliability based on non-probabilistic multi-ellipsoid models for improving the reliability and safety of transfer nursing robot. Firstly, the measurement of multi-source uncertainties is developed based on clustering ellipsoid for transfer nursing robot considering the distribution diversity. Secondly, the propagation model of multi-source uncertainties is established based on multi-ellipsoid pseudo probability to obtain the probabilities of response values for transfer nursing robot. Thirdly, the non-probabilistic reliability model considering multiple failure modes is constructed and the double-layer nested solution is developed. Finally, the proposed methods and models are integrated into a system and these research achievements are validated using the reliability analysis and experiments of double-arms transfer nursing robot. The completion of the projection will provide a novel research idea and technological approach for the propagation and reliability analysis under multi-source uncertainties.

安全可靠的移乘护理机器人对缓解护理人员紧缺、减轻子女负担、提高老人生活质量具有重要作用，应用前景广阔。针对移乘护理机器人在多源不确定性传播与可靠性分析中，因实验数据匮乏、变量分布多样、失效模式复杂以及双层嵌套求解困难所带来的建模难度大、计算效率差等瓶颈问题，本项目将基于非概率多椭球模型，发展适用于移乘护理机器人的多源不确定性传播分析方法和可靠性求解策略，以提高系统的可靠性与安全性。首先，基于聚类椭球发展考虑变量分布多样的移乘护理机器人多源不确定性度量方法；其次，建立基于多椭球伪概率的多源不确定性传播模型以获取系统响应任意取值的可能度；再次，构建考虑多失效模式的非概率可靠性模型并发展双层嵌套求解策略；最后，将相关方法和模型进行系统集成，通过双臂移乘护理机器人可靠性分析与实验，验证本项目的研究成果。本项目将为多源不确定性条件下移乘护理机器人的传播建模与可靠性分析提供一种新的研究思路和技术手段。

随着全球老龄化人口的日趋加剧，老人护理必将面临人手紧缺的难题。开发安全可靠的护理机器人，尤其是移乘护理机器人，对缓解护理人员紧缺、减轻子女及护理人员负担、提高老人生活质量等方面具有重要作用和应用前景。当前，移乘护理机器人的研究相对滞后，其设计通常是建立在确定性理论的基础上，采用设计参数的均值、极值等确定性参数值，而忽略参数的不确定性，这导致所设计的移乘护理机器人因对参数过于敏感，即便小的波动或差异都会造成严重的故障或伤害，从而影响移乘护理机器人的可靠性和安全性。在不确定性条件下，如何提高机器人的可靠性并确保人的安全性和舒适性是包括移乘护理机器人在内所有服务机器人以及协同机器人面临的共性问题。.本项目在非概率凸集模型理论框架下研究移乘护理机器人多源不确定性度量和可靠性分析方法，弥补随机概率法因样本信息匮乏而无法准确估计变量概率密度函数的缺陷，重点解决了采用传统非概率凸集模型方法面临由变量分布多样、失效模式复杂及双层嵌套求解带来的建模难度大、计算效率低等若干瓶颈问题。主要研究基于椭球可能度的移乘护理机器人多源不确定性度量、基于拉格朗日乘子与椭球体积比相结合的不确定性传播分析、考虑多失效模式的非概率可靠性建模与求解等内容，实现了移乘护理机器人变量高维相关、分布多样的不确定性传播分析与可靠性高效求解。此外，本项目的研究成果可以扩展至其他机器人或复杂装备的多源不确定性度量和可靠性分析，为相关领域高性能、高可靠性产品的开发提供理论参考和技术借鉴。