透過您的圖書館登入
IP:3.145.93.221
  • 學位論文

基於影像技術之棋藝機器人設計

Design of An Entertainment Robot Based on Image Processing Approach

指導教授 : 許駿飛

摘要


本論文提出了兩款四子棋機器人,分別為垂直式與水平式兩種不同的玩法,在設計上均使用了Raspberry Pi 3開發板與TM4C123GXL開發板,分別來做影像辨識以及遊戲人工智慧演算法等工作,並搭配網路攝影機、步進馬達、RC伺服馬達以及極限開關等工具來賦予機器人視覺以及執行下棋動作的能力。為了驗證所設計的兩款四子棋機器人成果,本論文提出合適的遊戲人工智慧演算法,讓機器人利用影像辨識模組去辨別玩家的下棋位置後,可以依據現有的棋盤狀況思考勝率最大的棋步,使機器人方的棋局保持優勢,同時讓玩家的棋局處於劣勢,藉此達到雙方互相對弈之娛樂效果。經由實驗結果可以發現,本論文所設計的遊戲人工智慧演算法能夠與玩家互相對弈,達到良好的消遣娛樂效果。

並列摘要


This thesis presents two connect-four robots which are vertical style and horizontal style respectively. Both of them use Raspberry Pi 3 LaunchPad and TM4C123GXL LaunchPad in the design. One launchpad is used to achieve image recognition module and the other one is used to implement the game artificial intelligence. They also use some tools like webcams, stepper motors, RC servo motors and limit switches to give the robot vision and the ability of placing a piece. In order to verify the effectiveness of the development interactive entertainment robot, this thesis designs two different connect-four games. With the game artificial intelligence, the robot use the image recognition module to recognize the player’s piece position and then thinking the maximum winning percentage in accordance with the board status. From the real-time result, it shows that the two connect-four robots can play with a player to achieve the effect of entertainment.

參考文獻


[10]甘禮豪,設計與實現互動式四子棋機器人,淡江大學電機工程學系碩士論文(指導教授:許駿飛),2012年。
[14]陳威宇,基於影像之機械手臂繪圖系統實現,淡江大學電機工程學系碩士論文(指導教授:許駿飛),2016年。
[1]L. George and S. William, Artificial Intelligence: Structures and Strategies for Complex Problem Solving, Benjamin-Cummings Publishing Company, 2004.
[2]K. Ray, The Singularity is Near, Penguin Books, 2005.
[3][Online]https://deepmind.com/research/alphago/

延伸閱讀