發展競賽型機器人的智慧所需要的知識領域相當的廣泛，必須結合累積相當經驗的領域專家共同的努力。本論文將競賽型機器人智慧的核心區分為「傳動」、「知覺」及「認知」三個部分。由知覺透過感應器感知的途徑獲取競賽場內的訊號，經過分析粹取出有用的訊息。認知將針對此訊息尋求合適的因應對策，決定如何妥善的回應。傳動則會命令驅動器執行回應，就競賽場內該情境實際的反應出本能行為。這過程均須由機器人自主地完成。
本研究以MSRDS開發工具賦予相撲機器人智慧，結合樂高機器人套件為硬體，利用恰當的傳動機構與靜態結構及創意巧思組裝二輪差動驅動相撲機器人。整個開發過程以AMDD(靈活塑模驅動開發，Agile Modeling-Driven Development)和UML(統一塑模語言，Unified Modeling Language)，按照「定義需求」、「分析及概念性設計」及「細節的設計」步驟進行塑模分析與設計。
本論文在使用VPL(視覺化程式語言，Visual Programming Language)實作對戰策略之前，先在VSE(視覺化模擬環境，Visual Simulation Environment)環境中進行模擬。當敵弱我強時，選擇追擊模式。若敵強我弱時，則採用迂回轉進模式。敵我強弱不明時，則攻守皆備。

To develop the contest robot’s brain must be more knowledge that conbine the different background, it needs to join some experts’s effect. In the thesis, the core of contest robot is “actuation”, “perception”, and “cognition”. The perception analyzes some useful messages that are got by sensor detecting for robot. The cognition will find the soiutions for these messages and make a decision how to response fittingly. The actuation will order the drivers to actuate some essential behaviors for the scenario in the competition field. These processes must be done by robot autonomy.
In the thesis, to use the development tool MSRDS to create the robot’s brain, and to assembly the suitable mechanism and structure using the lego robot parts for the 2WD SUMO robot. To use AMDD(Agile Modeling-Driven Development) and UML(Unified Modeling Language) in all the creation procedure, and modeling by step “requirement definition”, “analysis with comception design”, and “detail design”.
In this thesis, used to simulate in VSE(visual simulation environment) before creating a VPL(visual programming language) diagram to implement the strategies. When the enemy is weak and I am strong, choose to pursue and attack the way. When the enemy is strong and I am weak, adopt and transfer circuitously to. When enemy and we have unidentified power, offend and defend all fully.

[1]S. Lee, I.H. Suh, and M.S. Kim(Eds.) , Recent Progress in Robotics, LNCIS 370, pp. 385-397, (2008)
[2]Kevin Claycomb, Sean Cohen, Jacob Collums, Carlo Francis, Grzegorz Cieslewski, Tim Feeney, Donald J. Burnette, Gene Shokes, David Conrad, Hector Martinez III, Eric M. Schwartz, “SubjuGator 2007”, Florida Univ. (MIL), (2007)
[3]Anand Atreya ’07, Bryan Cattle ’07, Safiyy Momen ’07, Brendan Collins ’08, Alex Downey ’08, Gordon Franken ’08, Jon Glass ’08, Zach Glass ’08, Josh Herbach ’08, Andrew Saxe ’08, Issa Aswash ’09, Chris Baldassano ’09, Will Hu ’09, Umer Javed ’09, Jonathan Mayer ’09, David Benjamin ’10, Lindsay Gorman ’10, Derrick Yu ’10, “DARPA Urban Challenge Priceton University Technical Paper”, Priceton Univ. , pp.2-5, (2007)
[4]Scott W. Ambler, 朱子傑/陳盈學 譯，「THE OBJECT PRIMER 3/e 中文版—靈活模型驅動開發與UML 2」，碁峯，p4-18(2006)
[5]詹朝基、呂可望，「以本體論為基礎的知識庫發展架構與部落格網站設計實作」，國立台灣科技大學自動化及控制研究所碩士學位論文，pp.71-75，(2009)
[6]維基百科，http://zh.wikipedia.org/zh-tw/機器人。
[7]Defination of contract, internal state, behaviors and excution context，Kyle Johns, and Trevor Taylor, Professional Microsoft® Robotics Developer Studio, WROX, p85, (2008)
[8]James Floyd Kelly, LEGO MINDSTORMS NXT-G Programming Guide, APRESS Corp. (2007)
[9]Sara Morgan，康仕仲/蔡宛庭/涂智超/古凱元 編譯，「Microsoft® Roboics Studio程式開發—新一代超智慧機器人開發平台」，悅知文化，(2008)
[10]Microsoft, Microsoft Robotics Developer Studio 2008, Microsoft Corp. (2008)
[11]Kyle Johns, and Trevor Taylor, Professional Microsoft® Robotics Developer Studio, WROX, (2008)
[12]孫武，語農編注，「孫子兵法」，國家，(2008)
[13]台灣玉山機器人協會，http://www.era.org.tw/activities.php