擬人型機器手臂之機構設計與控制

Abstract

由於現在機器人越來越普及了，在日本，已經研究有十幾年的歷史，各項技術已漸趨成熟，而在台灣，專注於機器人研究的並不多。因此，在這方面是必須要趕快趕上的。 本文之主要目的，研究著重在機器手臂的機構設計與控制部份，所設計之機器手臂具有七個自由度，並且依照人的身體比例製作，使其運動可以像人類一樣，並期望未來可以裝在約150-160 cm 高的機器人身上。在控制方面，著重於硬體控制電路，並使用線上演算法，以達到多馬達之即時控制。 本論文可分為機構設計，運動學與動力學的分析，以及機器手臂的軌跡控制還有馬達控制。在機構設計部分，以繪圖軟體CATIA設計機構，以及工程設計軟體ADAMS去製作並且模擬運動學與動力學還有馬達所需要的扭力等，進而驗證實驗結果；並利用DSP (Digital Signal Processor) 單晶片以及FPGA (Field Programmable Gate Array) 去做馬達控制，並且與電腦端連接，再利用Matlab軟體去撰寫演算法與控制軌跡，並且探討奇異點、角度限制和障礙物的避開之演算法，使得機器手臂能夠依我們的指令，到達預期的位置，並且不會出現不穩定以及可以避開障礙物。還有力量控制部份，會利用力量感測器，來避免在末端夾爪作動時，可以依照所傳回來的力量，控制位置進行抓握的動作。

No doubt, robots have become more and more popular, and therefore, almost every technique on this area has already ripen up to now. Despite the fact that, the research of the humanoid robot has developed over more than ten years in Japan, currently only a few researchers in Taiwan devote themselves in this field. Hence, we should catch up quickly. The main purpose of this thesis focused on the mechanical design and control of the humanoid robot arm. The humanoid robot arm in question is about 150 to 160 centimeters in length, resembling a real human’s arm, with seven degrees of freedom. In control system, it is focused on the control hardware especially. In order to achieve the real-time multi-joint control, on-line algorithms are used. The discussions are divided into four parts: mechanical design of the arm, kinematics and dynamic analysis, trajectory planning, and hardware system control. In mechanical design, the CATIA software is used to design the mechanism, and ADAMS software is used to calculate and prove the motor torque and max speed and then finally verify the experimental result. As for the control system, we utilize the DSP (Digital Signal Processor) and FPGA (Field Programmable Gate Array) to do the motor control, and by connecting it to PC, and Matlab software to write high level algorithms, such as trajectory planning. In addition, the problem of redundant control to avoiding the joint angle limit, singularity, and obstacle will be discussed so that the robot arm will follow our instruction accurately. The FSR (Force Sensing Resistor) will be attached to the gripper, and it can grab objects sensitively.