Modular reconfigurable machines: theory and application

模块化可重构机器（MRM）代表了机器历史上的范式转变，因为它们能够根据环境和任务要求重新排列其模块并改变其结构。尽管世界各地已经开发了许多MRM，但它们仍然通过以前为传统机器开发的方法进行控制，但最近根据具体情况调整为MRM。当MRM改变其配置时，必须重新推导所有方程，包括运动学、动力学、运动静力学和校准。因此，新的理论必须发展，以处理不同的拓扑性质的MRM。在这项拟议的研究中，目标是开发一个基于模块的理论，有效的分析和控制的MRM。这项研究包括五个组成部分。第一部分是模块建模，其中MRM的模块将根据运动轴的数量和组合进行分类，然后将开发一种方法来对相邻模块之间的连接进行建模。第二部分是运动建模，其中模块固定坐标系将用于每种类型的模块，零参考平面方法将用于统一相对于全局坐标系的所有模块的运动表示。第三部分是系统建模，其中将在相邻模块之间开发转换函数，包括姿势/扭曲（运动），扳手（力），变形和误差，然后将这些转换函数嵌入到模块功能块中，这些功能块将被创建以自动生成用于MRM的各种分析和控制的系统方程。第四个组成部分是配置优化，其中枚举方法和基于遗传算法（GA）的方法将被开发用于搜索给定任务要求的MRM的最佳配置。最后一个组成部分是工业应用，其中两种类型的MRM将被研究，一个是模块化可重构机床（MRMT）和其他模块化可重构机器人（MRR），与制造业和航空航天的应用。

Modular Reconfigurable Machines (MRMs) represent a paradigm shift in the history of machines, because they are able to rearrange their modules and change their structures according to environment and task requirements. Although a number of MRMs have been developed around the world, they are still controlled by methods previously developed for traditional machines, but recently adapted to MRMs on a case-by-case basis. When the MRM changes its configuration, all equations, including kinematics, dynamics, kinetostatics, and calibration, must be derived all over again. Therefore, new theory must be developed in order to deal with the varying topological nature of MRMs. In this proposed research, the objective is to develop a module-based theory for effective analysis and control of MRMs. This research includes five components. The first component is module modeling in which the modules of MRMs will be classified in terms of the number and combinations of motion axes, and then a method will be developed to model the connectivity between adjacent modules. The second component is kinematic modeling in which a module-fixed coordinate frame will be used for each type of modules, and a zero-reference plane method will be used to unify the kinematic representation of all modules with respect to the global coordinate frames. The third component is system modeling in which transformation functions will be developed between adjacent modules, including pose/twist (motion), wrench (force), deformation and error, and then these transformation functions will be embedded in the module function blocks, which will be created to automatically generate the system equations for various analyses and controls of MRMs. The fourth component is configuration optimization in which both an enumeration method and Generic Algorithms (GA)-based method will be developed to search for an optimal configuration of a MRM for given task requirements. The last component is industrial applications in which two types of MRMs will be studied, one being modular reconfigurable machine tools (MRMTs) and the other being modular reconfigurable robots (MRRs), with applications to manufacturing and aerospace.