透過您的圖書館登入
IP:3.21.100.34
  • 學位論文

多群組機器人系統中資訊交換與隊形控制互動之分析

Analysis of Interaction between Information Exchanging and Formation Control in Multi-Robot Systems

指導教授 : 連豊力

摘要


近年來,多群組機器人系統已逐漸被廣泛研究。由一群功能簡單的機器人所組成的群組可借由合作行為來取代單一而較複雜的機器人。在合作的過程中,機器人若能形成特殊的幾何隊形,通常有助於提高工作及彼此間資訊交換的效率。 為了要形成特殊的幾何隊形,機器人之間的互動應該要被妥善地安排。這些互動可以分為兩類:行動互動以及資訊互動。行動互動指的是機器人之間的相互運動而資訊互動則意指機器人間交換感測器讀數或通訊資料。在本文中,機器人在空間上的相互關係以及資訊交換時所形成的資訊拓樸可用基本的圖學理論來表示。幾何隊形的穩定度可借由非線性系統分析中有關系統穩定度的概念來加以定義。控制機器人行動互動的高層級控制器則使用人造位能場方法來設計,此控制器可驅使機器人群排成一個穩定的幾何隊形。考慮到不同資訊拓樸對於行動互動的影響,在本文中,我們討論了五個不同的例子。另外,在本文中使用了如強健性、收斂速度、能量消耗、敏感度以及整體系統效率等五個不同的指標來比較各個不同資訊拓樸的優缺點。為了能使機器人群組改變資訊拓樸,一個簡單而實用的控制器切換演算法在本文中被提出。五個不同例子的模擬結果顯示若資訊拓樸選擇得好,在形成特殊幾何隊形時,機器人群組的行動互動可以更有效率且更強健。

並列摘要


The research topics of multi-robot systems have attracted much attention in recent years. A collection of simple robots can replace a complex robot via cooperation. In the process of cooperation, forming a specific formation pattern will enhance the efficiency of work and data transmission between robots. In order to form a specific formation pattern, interaction between robots should be well regulated. The interaction can be divided into two groups: motion interaction and information interaction. Motion interaction means the relative movement between robots and information interaction means the exchanges of sensing readings or communication data of robots. In this thesis, graph theory is used to model the physical relationship and information exchanging topology among robots. Stability of formations is defined according to the stability concept in nonlinear system analysis. Artificial potential field based high-level controllers are designed for the motion interaction of robots to form stable formations. Five cases with six robots in formation control problem considering effects of different information topologies are presented and discussed. Factors such as robustness, convergence speed, power consuming, sensitivity, and system efficiencies are compared for different information topologies between six robots. A simple controller switch algorithm for switching between two different information topologies is proposed. Simulation studies of the five cases show results that motion interaction of robots in formation control can be more efficient and robust with properly chosen information patterns.

參考文獻


[1: Khalil 2002]
[2: Siegwart and Nourbakhsh 2003]
[3: Murphy 2000]
Graph theory:
[4: Krishnamurthy and Preis 2005]

延伸閱讀