透過您的圖書館登入
IP:3.135.216.174
  • 期刊

動作辨識及模仿在人型機器人之實現

Implementation Motion Recognition and Imitate in Humanoid Robot

摘要


本論文的目的在控制具視覺系統之人型機器人對肢體動作的辨識與模仿。人型機器人身體上的關節是由17顆AI馬達加以驅動,手部以及足部的關節是由16顆AI馬達所構成,頭部由一顆AI馬達與馬達上之無線CCD(Charge Coupled Device)結合而成,背部之控制面板可用於控制所有AI馬達之角度方位,使機器人能夠做出各種不同動作。系統架構上使用無線CCD擷取影像,其接收器透過影像擷取卡UPG304將影像傳送至PC端進行影像處理。經過影像分析後,透過本文所提出的模仿動作編成可取得模仿動作之控制命令,再經由ZigBee無線傳輸模組將控制命令傳輸至機器人的控制背板,驅動AI馬達完成模仿動作。

並列摘要


In this thesis, the control and image processing algorithm is proposed to a humanoid robot with visual system, so that it can recognize and imitate the motion of human. The hardware architecture of humanoid robot is mainly composed of 17 AI motor including the head of the wireless CCD (Charge Coupled Device), and control board. A wireless CCD module with video capture card UPG304 is used to transmit the image to PC. The image processing algorithm is designed to analyze and recognize the motion of arms or legs, then the properly control command is generated and transmitted through the ZigBee wireless module to humanoid robot to imitate the recognized motion.

延伸閱讀