結合網路與虛擬實境之遠端機器人系統：阻抗性控制與視覺系統

Abstract

在一些未知的或是具有危險性的環境中，我們可以使用機器人與遠端操控技巧來幫助我們完成任務。一方面我們藉由虛擬實境技術來提供遠端環境三維場景的視覺感受，另一方面，我們藉由裝設在機器人末端的力感測器來偵測機器人與環境接觸時所產生的接觸力，並回授至操作端使我們同時也能感受到操作過程的接觸力。此外，當我們操作機器人來執行順應性工作時，機器人末端點位置以及機器人與環境所產生的接觸力兩者之間的互動變得更為重要，所以系統必須擁有更有效的順應性控制技術。因此在本篇論文中，我們針對系統所使用的位置伺服控制機器人，運用兩種適應性阻抗控制法則，在不須要知道環境的位置與硬度的情況下，得到機器人的參考位置軌跡以達到所指定的接觸力。另外，視覺系統被也納入此遙控機器人系統來完成真實環境中物體的定位以及識別。我們也透過幾項實驗來驗証控制策略的可行性，以及系統執行順應性工作的能力。

For hazardous or uncertain environments, such as the nuclear power plants or the outer space, the robot combined with the teleoperation technique is helpful for the human in task execution. For the proposed telerobotic system, virtual reality technology is introduced into the system to provide 3D visual display of the remote environment. In addition, the contact force between the robot and the environment is sensed by the force sensor mounted on the robot's end-effector, and fedback to the master site. Therefore, the operator can also feel the interactive force during teleoperation. Furthermore, when we manipulate the robot to execute a compliance task, the treatment of the interaction between the robot and environment is very important, so that there must be some effective compliance control schemes in the system. In this thesis, two adaptive impedance control schemes are introduced for the position-controlled robot in our system. By applying these impedance control schemes, the desired contact force can be achieved by the generated reference position trajectory for the robot manipulator, without the information of environmental location and stiffness. In addition, a vision system is incorporated into the system to accomplish the localization and identification of the objects in the real environment. We also evaluate system performance through several experiments. The results demonstrate the feasibility of the control schemes and the capability of the system in compliance task execution.