對機器人足球的決策與路徑規(guī)劃的研究
發(fā)布時間：2018-12-12 22:05
【摘要】： 機器人足球比賽是近年來在國際上迅速開展起來的高技術對抗活動。它是人工智能領域與機器人領域的基礎研究課題，是一個極富挑戰(zhàn)性的高技術密集型項目。本文以足球機器人系統(tǒng)的核心子系統(tǒng)——決策子系統(tǒng)的設計開發(fā)為背景，研究行之有效的決策推理方法。足球機器人系統(tǒng)是一個典型的多智能體系統(tǒng)，涉及機器人學、計算機視覺與模式識別、多智能體系統(tǒng)、軌跡規(guī)劃與智能算法、自組織與自學習理論等領域。 足球機器人系統(tǒng)分為四個子系統(tǒng)——機器人子系統(tǒng)、視覺子系統(tǒng)、決策子系統(tǒng)和通訊子系統(tǒng)。其中決策子系統(tǒng)是整個系統(tǒng)的核心，應具有可自主完成知識提取，，并確定機器人協(xié)作任務的能力，使整個系統(tǒng)具有智能體的特征。機器人足球環(huán)境是一個具有動態(tài)性、不確定性、實時性的環(huán)境，在這樣一個具有高度實時性和競爭性平臺上研究路徑規(guī)劃也是一個很具有挑戰(zhàn)性的課題。通過本課題的研究，得到如下的成果與結論： 1．綜合國內(nèi)外相關研究文獻，闡述了足球機器人研究現(xiàn)狀和主要內(nèi)容。同時，詳細介紹了足球機器人系統(tǒng)、路徑規(guī)劃及機器人隊形確定問題的研究現(xiàn)狀。 2．介紹了幾種足球機器人路徑規(guī)劃的理論和算法，，如人工勢場法、柵格法、可視圖法以及各種人工智能方法如遺傳算法、神經(jīng)網(wǎng)絡等。但這些方法在高度動態(tài)和實時的環(huán)境中的研究還不太完善，需要進一步改進。討論它們在足球機器人系統(tǒng)中的應用可行性，最終確定以人工勢場法為研究方向。 3．通過對足球機器人隊形確定問題的研究，來闡述足球機器人路徑規(guī)劃中的路徑?jīng)_突協(xié)調(diào)和協(xié)作問題。 4．通過一種新的開源足球機器人仿真平臺robotsoccer建立仿真實驗環(huán)境，然后在在平臺上利用lingo編程語言編寫客戶端算法，實現(xiàn)對仿真足球機器人的控制。 本論文在對足球機器人總體系統(tǒng)進行介紹的基礎上，分析研究了決策子系統(tǒng)的系統(tǒng)模型，提出了一種基于改進人工勢場法進行動態(tài)環(huán)境的路徑規(guī)劃的方法。通過梯度逼近進行運動目標的搜索，解決了在實時路徑規(guī)劃中因環(huán)境的運動信息難以準確獲取而造成的路徑規(guī)劃無法完成的問題。仿真結果驗證了方法的有效性，能夠較好的解決動態(tài)環(huán)境下特別是存在隨機運動物體情況下足球機器人的路徑規(guī)劃問題。
[Abstract]:Robot soccer match is a high-tech counteraction that has been developed rapidly in the world in recent years. It is a basic research subject in the field of artificial intelligence and robot. It is a challenging high-tech-intensive project. Based on the design and development of the decision-making subsystem, which is the core subsystem of the soccer robot system, an effective decision reasoning method is studied in this paper. Soccer robot system is a typical multi-agent system, which involves robotics, computer vision and pattern recognition, multi-agent system, trajectory planning and intelligent algorithm, self-organization and self-learning theory and so on. Soccer robot system is divided into four subsystems-robot subsystem, vision subsystem, decision-making subsystem and communication subsystem. The decision-making subsystem is the core of the whole system, and it should have the ability to complete the knowledge extraction independently and determine the cooperative task of the robot, so that the whole system has the characteristics of the agent. Robot soccer environment is a dynamic, uncertain and real-time environment, and it is also a challenging task to study path planning on such a highly real-time and competitive platform. Through the research of this topic, the following results and conclusions are obtained: 1. The current situation and main contents of soccer robot research are expounded by synthesizing the relevant research literature at home and abroad. 2. At the same time, the research status of soccer robot system, path planning and robot formation determination is introduced in detail. 2. Several theories and algorithms of soccer robot path planning are introduced, such as artificial potential field method, grid method, visual graph method and various artificial intelligence methods such as genetic algorithm, neural network and so on. However, the research of these methods in high dynamic and real-time environment is not perfect, which needs further improvement. The feasibility of their application in the soccer robot system is discussed, and the artificial potential field method is finally chosen as the research direction. 3. The coordination and cooperation of path conflict in soccer robot path planning are discussed by studying the problem of soccer robot formation determination. 4. The simulation experiment environment is established by a new open source soccer robot simulation platform, robotsoccer, and then the client algorithm is programmed by lingo programming language on the platform to realize the control of the simulation soccer robot. Based on the introduction of the overall system of soccer robot, the system model of decision-making subsystem is analyzed and studied in this paper, and a path planning method based on the improved artificial potential field method is proposed. By using gradient approximation to search moving objects, the problem that path planning can not be completed in real-time path planning is solved because the environment motion information is difficult to obtain accurately. The simulation results show that the method is effective and can solve the path planning problem of soccer robot in dynamic environment, especially in the case of random moving objects.
【學位授予單位】：山東大學
【學位級別】：碩士
【學位授予年份】：2007
【分類號】：TP242
本文編號：2375308
[Abstract]:Robot soccer match is a high-tech counteraction that has been developed rapidly in the world in recent years. It is a basic research subject in the field of artificial intelligence and robot. It is a challenging high-tech-intensive project. Based on the design and development of the decision-making subsystem, which is the core subsystem of the soccer robot system, an effective decision reasoning method is studied in this paper. Soccer robot system is a typical multi-agent system, which involves robotics, computer vision and pattern recognition, multi-agent system, trajectory planning and intelligent algorithm, self-organization and self-learning theory and so on. Soccer robot system is divided into four subsystems-robot subsystem, vision subsystem, decision-making subsystem and communication subsystem. The decision-making subsystem is the core of the whole system, and it should have the ability to complete the knowledge extraction independently and determine the cooperative task of the robot, so that the whole system has the characteristics of the agent. Robot soccer environment is a dynamic, uncertain and real-time environment, and it is also a challenging task to study path planning on such a highly real-time and competitive platform. Through the research of this topic, the following results and conclusions are obtained: 1. The current situation and main contents of soccer robot research are expounded by synthesizing the relevant research literature at home and abroad. 2. At the same time, the research status of soccer robot system, path planning and robot formation determination is introduced in detail. 2. Several theories and algorithms of soccer robot path planning are introduced, such as artificial potential field method, grid method, visual graph method and various artificial intelligence methods such as genetic algorithm, neural network and so on. However, the research of these methods in high dynamic and real-time environment is not perfect, which needs further improvement. The feasibility of their application in the soccer robot system is discussed, and the artificial potential field method is finally chosen as the research direction. 3. The coordination and cooperation of path conflict in soccer robot path planning are discussed by studying the problem of soccer robot formation determination. 4. The simulation experiment environment is established by a new open source soccer robot simulation platform, robotsoccer, and then the client algorithm is programmed by lingo programming language on the platform to realize the control of the simulation soccer robot. Based on the introduction of the overall system of soccer robot, the system model of decision-making subsystem is analyzed and studied in this paper, and a path planning method based on the improved artificial potential field method is proposed. By using gradient approximation to search moving objects, the problem that path planning can not be completed in real-time path planning is solved because the environment motion information is difficult to obtain accurately. The simulation results show that the method is effective and can solve the path planning problem of soccer robot in dynamic environment, especially in the case of random moving objects.
【學位授予單位】：山東大學
【學位級別】：碩士
【學位授予年份】：2007
【分類號】：TP242
【引證文獻】
相關期刊論文 前1條
1 胡玲;廖家平;舒軍;魯海霞;;足球機器人的路徑規(guī)劃方法[J];天津市經(jīng)理學院學報;2010年04期
相關碩士學位論文 前1條
1 伍龍軍;MiroSot足球機器人決策子系統(tǒng)的研究[D];西華大學;2008年
本文編號：2375308
本文鏈接：http://sikaile.net/kejilunwen/rengongzhinen/2375308.html
教材專著
