项目名称: 虚拟人维修动作混合驱动及人机工效自动量化评估方法研究
项目编号: No.51475291
项目类型: 面上项目
立项/批准年度: 2015
项目学科: 机械、仪表工业
项目作者: 何其昌
作者单位: 上海交通大学
项目金额: 80万元
中文摘要: 针对现有虚拟维修系统中维修指令人机工效评估效率低下、准确度不高,不能满足复杂装备维修性并行设计的需求,本课题建立维修工具信息模型,研究基于全局可接近性锥GAC的维修工具最优匹配方法,提高维修工具的匹配效率;分析虚拟人驱动精度影响因素,研究基于遗传算法的虚拟人运动精度优化控制技术,提高捕捉数据动作驱动的精度;研究基于功能约束的精细维修动作建模技术,建立虚拟人姿态生成的多层次约束模型,实现精细维修动作的算法驱动;研究维修动作分类方法,建立基于实时捕捉数据与模型混合驱动的虚拟人运动控制机制,生成准确的虚拟人动作姿态。研究基于核聚类的关键帧自动提取技术与人机工效信息的工程量化准则及计算方法,形成面向维修过程的维修指令人机工效自动量化评估的技术体系;通过多样本的实验设计,校验人机工效量化结果的可信性;最后将集成研究成果,开发软件原型系统,并通过飞机、雷达等复杂装备的应用对方法的性能进行分析。
中文关键词: 虚拟维修;虚拟人;混合驱动;人机工效;自动量化评估
英文摘要: The human factors evaluation in current virtual maintenance system is low efficency and accuracy which can not satisfy the requirements of concurrent maintenance design of complex equipment. Facing these problems, this study research on information modeling of maintenance tools oriented to the virtual simulation after analyzing typical maintenance order. Then develop a hybrid driven model of maitenance action of virtual human, which combining the real-time interactive control method and Parameterized Action Representation method, to realize the simulation driven from maintenance instruction written by natural language to 3D maintenance simulation, so as to improve the manual simulation setting mode. At the same time, the study develops the automatic quantitive evaluation technology for human factors based on the real-time virtual maintenance simulation process. Finally, this study develops a prototype system after integrating research achievements and using radar,plane equipment to verify those methods.
英文关键词: virtual maintenance;virtual human;hybrid driven;human factors;automatic quantitive evaluation