多機器人合作感測定位與環境地圖建立

Abstract

環境地圖與自我定位是智慧型機器人導航的重要資訊。最近幾年，利用多機器人合作感測定位的研究逐漸受到重視，如何提供智慧型機器人對本身位置做準確的估測，進而建立可供後續使用的環境地圖是此研究領域的重點。本論文發展出一個運用多機器人系統以合作式感測並利用卡曼濾波器概念來對機器人團隊中的機器人做感測資訊融合，降低其位置不確定性，進而建立較準確的線段環境地圖。本論文建立一套多機器人Server-Client架構，用來對多機器人系統做一有效的操作與監控。最後利用2台機器人做合作感測定位的驗證，結果顯示所提出之方法確實有預期的功效。

Map-building and self-localization are important for intelligent mobile robots. In recent years, the problems in multi-robot cooperative sensing have been an interesting research area. The focus of this thesis is simultaneous localization and map-building based on multi-robot cooperation. This thesis presents a method to localize robots, and construct a two-dimensional(2D) map based on stereo vision from multiple mobile robots. Sensed information by multiple robots is fused using Kalman filtering techniques to reduce the uncertainty of robot positions. We setup a client-server architecture to coordinate robot tasks. Our method has been demonstrated with experiments using two mobile robots. Satisfactory experimental results are presented in the thesis.