Modular Distributed Manipulator System

本研究的重点是一种新型材料运输和机械手系统的控制和运动规划算法，称为模块化分布式机械手系统（MDMS），其应用范围从柔性制造到包装处理。MDMS包括一组致动器，其中每个致动器可向其上的物体施加定向力。该项目开发了测试平台硬件和算法来协调执行器来传输和操纵对象。理论贡献增加了分布式操作的现有文献，其中大部分涉及芯片级执行器。MDMS在宏观尺度上工作，物理模型更容易处理。与MEMS致动器阵列相比，在MEMS致动器阵列中，质量和摩擦基本上被忽略了，MDMS不仅包含了摩擦和质量，而且还明确地利用了它们。这项工作考虑了两种主要的操作模式：无传感器和基于传感器。无传感器操作更容易实现和分析，但不提供基于传感器的控制的精度。MDMS还为评估分布式控制策略提供了一个测试平台，并且本研究考虑了集中和分散控制方法之间的权衡。这些算法将在现有的18单元原型上进行演示。一个更大、更先进的原型机将被开发出来进一步测试这一概念。

This research focuses on the control and motion planning algorithms for a new materials transport and manipulator system called the Modular Distributed Manipulator System (MDMS), which has applications ranging from flexible manufacturing to package handling. The MDMS comprises an array of actuators each of which can impart a directed force to an object resting upon it. This project develops testbed hardware and algorithms to coordinate the actuators to transport and manipulate objects. The theoretical contributions add to the existing literature on distributed manipulation, much of which addresses chip-scale actuators. The MDMS works on a macroscopic scale, where physical models are more tractable. In contrast to MEMS actuator arrays, where mass and friction are essentially ignored, the MDMS not only incorporates friction and mass, but also explicitly exploits them. This effort considers two primary operation modes: sensorless and sensor-based. Sensorless operation is easier to implement and analyze but does not provide the precision of sensor-based control. The MDMS also provides a testbed for evaluating distributed control strategies, and this research considers the tradeoff between centralized and decentralized control approaches. These algorithms will be demonstrated on an existing eighteen cell prototype. A larger, more advanced prototype will be developed to further test the concept.