https://scholars.lib.ntu.edu.tw/handle/123456789/152160
標題: | 總計畫: 具多個互動型動態代理器之大型網狀混合式控制系 統 | 作者: | 傅立成 | 關鍵字: | 動態代理器;行動輪式機械人;網路聯界結;軟硬體測試平台;Dynamic Agent;Mobile manipulator;Network Linkage;Software and hardware testing platform | 公開日期: | 31-七月-2004 | 出版社: | 臺北市:國立臺灣大學電機工程學系暨研究所 | 摘要: | 一個動態系統的感測器及致動器傳統上均以專線(hard wire)透過某種型式A/D 及D/A 及個人電腦之CPU 或單晶片控制器(卡)相聯,其中所需之控制演算法泰半能以即時方式完 成運算,但由於網路技術日新月異、突飛猛進,該專線連結近來漸有以網路連結方式取而 代之,尤其是當需連結之單元數相當龐大時,上述取代已別無選擇。不過網路連結方式傳 統上欲達到即時反應的要求的困難度高,勢將造成控制效能上的負擔,不過直觀上只要網 路頻寬無限制提昇,則即時控制即不再是空談,但殊不知前者在目前現階段並不可能,而 此種限制亦可透過巧妙地重新設計控制器的架構、控制演算法、與控制邏輯來達到殊途同 歸的效果。本整合型計畫的目的即在探討與驗證上述的可行性。 子計畫一在本大型計畫中扮演中心引導、整合各其他子計畫成果的角色,目標係提供 各子計畫理論、技術測試、驗證的平台,當然初期亦可作為各子計畫執行的動機來源。此 平台的提創、開發與製作在此三年的計畫中將循序漸進。初期、本子計里將首先提出一個 具互動、動態子系統的大型網狀系統,為使其具相當代表性,將考慮以行動輪式機械人 (mobile manipulator)作為動態代理器,其中輪式機械人部分可作平面移動及旋轉,而機 械臂則為多軸,此代理器設有一局域性控制器(local controller),可控制車體與機械臂。 此等多個自走式機械臂將試圖合作將一大型物件合力移動,其中移動的軌跡可由一遙距的 控制器以網路連結的方式動態控制各個代理器的機械臂、抑或各代理器之間亦透過網路協 調彼此間的互動,藉由其間力的動態平衡達到各代理器及物件的穩定性。此系統的穩定性 明顯將繫於網路頻寬,是以大型網狀系統將完全符合整合型計畫所需。 在第一年提出並分析各代理器之架構後,將同時製作其prototype,第二年,本子計 畫將製作一大型軟體模擬系統以供各子計畫透過網路方式上線模擬測試,第三年則實際製 作多個代理器, 屆時各子計畫將實地驗證其所提出之理論。 Traditionally, the sensing subsystem and actuation subsystem are linked to the main CPU of the control computer or some single-board micro-controller through some hardwiring and A/D, D/A converters. As a result, the control algorithms can mostly be completed in real time. Recently, due to dramatic renewal and improvement of the network technology, it is becoming a trend to replace the above hardwire links by the network links, and it becomes an indispensable choice especially when the number of units to be linked is enormous. However, it is well known that the real-time response is hardly come by from the network linkage, which thereby causes burden in the overall control performance. Intuitively, the resolve the real-time control dilemma is to arbitrarily increase the bandwidth of the network. But the current technology, however, did not support this request, and in fact such a problem can be overcome though ingeniously redesign of the control architecture, the control algorithm and the control logic. This subproject is solely to investigate and to verify the possibility of the mentioned solution alternative. This subproject, among the present integrated project, is to play the role of guiding the research of and integrating all the results from other subprojects. Specifically, the goal of this subproject is to provide a platform to test and to assess the theoretical results and technological developments. Of course, this subproject can also have a role in highly motivating all the execution of other subprojects. The proposition and development of this platform will proceed in this three year integrated project step by step. In the beginning period, this subproject will propose a large scale networked system consisting of multiple interactive dynamic subsystems. An excellent choice is to choose a mobile manipulator as a dynamic agent, where the mobile vehicle can move in a plane and can rotate whereas the manipulator can have three joints, i.e., with three degree-of-freedoms (DOF’s). Furthermore, each agent will have its own local controller to control the vehicle and the manipulator. These many agents will then cooperate to move a huge piece of large object, and the motion trajectory of the object as well as of each agent will be determined by a remote controller through wireless network or be determined after negotiation among various relevant agents. It is crucial that this dynamic network communication will guarantee the force balance in a dynamic manner and in turn lead to full stability of the large object as well as each relevant agent. It is noteworthy that the stability of the overall system has a strong tie with the bandwidth, and therefore the present system indeed will fulfill the need by the present integrated project. After we have proposed the architecture of the agent in the first year, we will build the prototype at the end of the same year. Next, in the 2nd year we will develop a large scale simulation software to facilitate every subproject to assess its result on-line. On the 3rd year, we will develop multiple prototype agents, so that all the subprojects can evaluate their results performance experimentally and physically. |
URI: | http://ntur.lib.ntu.edu.tw//handle/246246/7980 | 其他識別: | 922213E002044 | Rights: | 國立臺灣大學電機工程學系暨研究所 |
顯示於: | 電機工程學系 |
檔案 | 描述 | 大小 | 格式 | |
---|---|---|---|---|
922213E002044.pdf | 792.41 kB | Adobe PDF | 檢視/開啟 |
在 IR 系統中的文件,除了特別指名其著作權條款之外,均受到著作權保護,並且保留所有的權利。