Reconfigurable Cable-Driven Parallel Robots

绳索驱动并联机器人(CDPR)由一个末端执行器组成，末端执行器由多根绳索悬挂。每根电缆都缠绕在 一种伺服驱动的卷轴，它牢固地固定在固定的框架上。通过同时缠绕和解开各自线轴上的缆线来控制末端效应器的位移。我们对CDPR的兴趣来自于它们相对于传统刚性连杆机器人的五个优势：(I)它们覆盖了很大的工作空间；(Ii)它们便宜；(Iii)它们可以产生高速和加速-据报道超音速！；(Iv)它们是便携的；(V)它们是可重新配置的。在我们看来，CDPR的最后一个优势，即快速重新配置的能力，在以前的研究中相对被忽视了。从许多实际应用的角度来看，这一点是至关重要的，因为机器人可以根据不同的工件或环境调整其几何形状和工作空间，而大多数传统机器人无法做到这一点。 因此，提出的研究目的是改进和展示缆索驱动并联机器人的重构能力。我们已经确定了工程师在为新任务重新配置CDPR时面临的三个重要挑战：为给定的任务找到最佳的机器人几何形状；更换末端执行器或其执行工具(例如，夹爪、焊枪)；以及在重新配置后确定准确的CDPR几何形状。拟议的研究计划将通过一系列从4.1到4.5的五个项目来应对这些挑战。 在项目4.1中，我们将开发一种新的方法来跟踪CDPR的三维工作空间。此工作空间将考虑缆索中的张力限制、它们将产生的合力和力矩的范围，以及缆索与末端执行器之间可能发生的碰撞。在项目4.2中，我们将开发一种自动合成CDPR的新方法。这种方法将考虑到项目4.1中包括的所有实际限制，这是目前可用的方法所不可能做到的。为了验证其有效性，将使用综合方法设计两个CDPR：一个用于大型曲面的检查(项目4.5)，另一个用于在虚拟环境中提供物理康复的触觉反馈。 在项目4.3和4.4中，我们将提高CDPR工具的互换性，将其执行器从末端执行器移动到地面，从而允许对不同的工具使用相同的电机和控制器。这将降低工具的单位成本，从而允许制造更多不同的工具。运动和力将通过与已形成CDPR的电缆平行的附加电缆从固定致动器传递到移动末端执行器。在项目4.3中，我们将通过在差动模式下使用两个电机来控制单自由度刀具。在项目4.4中，我们将把这一概念扩展到从底座驱动三自由度机械臂，从而改进了机器人学中众所周知的宏微型机械手的想法。 在项目4.5中，我们将应用之前开发的分析和综合方法来设计用于大型表面自动无损检测(NDT)的CDPR。探测器将安装在CDPR末端执行器上，并用于扫描各种几何形状的对象的模型，例如飞机机翼或储液器。该项目将使用项目4.1和4.2的结果，并将需要开发新的校准技术，以便在重新配置后可以准确地识别CDPR几何图形。 除了为五名学生提供最先进的培训外，这项研究还将把CDPR的重新配置能力提升到可以应用于工业任务的功能水平。

Cable-driven parallel robots (CDPRs) consist of an end effector, which is suspended by a number of cables. Each cable is wound onto a servo-actuated spool, which is rigidly attached to a fixed frame. The end effector displacements are controlled by simultaneously winding and unwinding the cables on their respective spools. Our interest in CDPRs stems from five advantages they hold over classical rigid-link robots: (i) They cover a large workspace; (ii) They are inexpensive; (iii) They can generate high speeds and accelerations---Supersonic speeds have been reported!; (iv) They are portable; (v) They are reconfigurable. In our opinion, this last advantage, the ability of CDPRs to quickly reconfigure, has been relatively neglected in previous research. From the perspective of many practical applications, this point is crucial, as the robot can then adapt its geometry and workspace to different workpieces or environments, which most classical robots cannot do. The objective of the proposed research is thus to improve and demonstrate the ability of cable-driven parallel robots to reconfigure. We have identified three important challenges facing the engineer when reconfiguring a CDPR for a new task: finding the best robot geometry for a given task; changing the end effector or its actuated tool (e.g., gripper, welding torch); and identifying the exact CDPR geometry after it has reconfigured. The proposed research program will address these challenges in a series of five projects numbered from 4.1 to 4.5. In project 4.1, we will develop a new method of tracing the three-dimensional workspaces of CDPRs. This workspace will take into account the tension limits in the cables, the range of resultant forces and moments they are to generate, and the possible collisions between cables and end effector. In project 4.2, we will develop a new method for the automatic synthesis of CDPRs. This method will take into account all the physical constraints included in project 4.1, which is not possible with currently available approaches. To demonstrate its effectiveness, the synthesis method will be used to design two CDPRs: one for the inspection of large surfaces (project 4.5), and the other, to provide haptic feedback in a virtual environment for physical rehabilitation. In projects 4.3 and 4.4, we will improve the CDPR tool interchangeability by moving its actuators from the end effector to the ground, thus allowing to use the same motors and controllers for different tools. This will reduce the unit cost of a tool, thus allowing for more different tools to be made. The motion and forces will be transmitted from the fixed actuators to the moving end effector via additional cables parallel to those already forming the CDPR. In project 4.3, we will control a tool with a single degree of freedom by using two motors in differential mode. In project 4.4, we will extend this concept to drive a three-degree-of-freedom robotic arm from the base, thus improving on the idea of the macro-mini manipulator, which is well known in robotics. In project 4.5, we will apply the analysis and synthesis methods previously developed to design a CDPR for the automatic non-destructive testing (NDT) of large surfaces. A probe will be mounted on the CDPR end effector, and used to scan mockups of objects of various geometries, e.g., an aircraft wing or a reservoir. This project will use the results of projects 4.1 and 4.2, and will require the development of new calibration techniques, so that the CDPR geometry can be precisely identified after it has reconfigured. Besides providing state-of-the-art training to five students, this research will take the reconfiguration capabilities of CDPRs to a functional level that can be applied to industrial tasks.