夜視型自主式群組校園巡邏機器人之研究--總計畫：夜視型自主式群組校園巡邏機器人之研究
No Thumbnail Available
Date
2011-07-31
Authors
王偉彥
Journal Title
Journal ISSN
Volume Title
Publisher
行政院國家科學委員會
Abstract
本整合型計畫的最終目標是讓群組機器人，能夠自主的在日間(一般CCD 辨識技術)、夜間或是光 線不佳(整合紅外線熱像儀影像辨識技術)的環境下，進行巡邏任務。子計畫二將以派屈網路理論，開 發夜間群組巡邏機器人之網路通訊、決策機制並以模糊派屈網路控制巡邏機器人之單機以實現分散式 網路控制概念。此外子計畫一將研發專屬群組機器人的資料庫、訊息整合系統與3D 使用者介面，建 立機器人估測電池電量及電池健康度之演算法。子計畫三將著重於發展未知環境之探索技術，藉各類 感測元件達到自主探索環境，完成機器人巡邏任務之環境地圖、地標標示與路徑規畫。子計畫四將利 用熱影像攝影發展非接觸式生物識別模組，另一方面，將根據熱影像攝影發展意外偵測模組(包含火災 識別模組與姿態模組)，此為使用熱影像式攝影機進行生物識別的附加價值。子計畫五使用熱影像攝影 技術使得機器人在巡視校園時，可以不受任何光線的影響，皆能清楚地偵測估算出環境中的行人與行 人的行徑方向和速度及相對位置。最後本整合型計畫將完成各項功能開發研究，以提高機器人在學界 和業界的實用性。
The main objective of this integrated project is to make possible for group robots to autonomously patrol around the campus during day and nighttime, as well as in weak lighting surroundings. These robots must be capable of patrolling in difficult campus locations, while at the same time detect pedestrians, animals, vehicles, and obstacles. In order to understand the concept of distributed network control, sub-project two will apply the Petri net theory, in wanting to build communications between robots and the decision-making mechanism. A design is also included in Sub-project two, which uses the control law to manipulate a single robot by applying the Fuzzy Petri net theory. In addition, the purpose of sub-project one is to design and build a database, a message integration system, and a 3D user interface for group robots. Sub-project one also derives an algorithm which estimates the battery capacity of robots and the battery condition by the hierarchical fuzzy-neural network. As sub-project three focuses on developing an approach to explore the unknown environment by using many sensors, it also evolves functions of drawing special maps for group robots after obtaining the information in an unexplored environment, orientating and making the motion planning. Next, sub-project four will develop a non-contact infrared thermal imager method for the biometric identification, and moreover develop a disaster detection module that includes fire and posture detections after getting access to required accident information. Lastly, sub-project five is a research done in a group of campus patrolling robots by using infrared thermal imager techniques, which is a detection process that avoids unwanted consequences caused by factors such as nighttime or weak lighting. In other words, it can hence successfully detect the pedestrian, the position, velocity, and the orientation of the pedestrian in the campus. To conclude, the research group will combine all research findings of patrolling group robots; one believes that the accomplishment of this integrated project can be a big step in the development of robots in both the academic research field, as well as in industrial applications.
The main objective of this integrated project is to make possible for group robots to autonomously patrol around the campus during day and nighttime, as well as in weak lighting surroundings. These robots must be capable of patrolling in difficult campus locations, while at the same time detect pedestrians, animals, vehicles, and obstacles. In order to understand the concept of distributed network control, sub-project two will apply the Petri net theory, in wanting to build communications between robots and the decision-making mechanism. A design is also included in Sub-project two, which uses the control law to manipulate a single robot by applying the Fuzzy Petri net theory. In addition, the purpose of sub-project one is to design and build a database, a message integration system, and a 3D user interface for group robots. Sub-project one also derives an algorithm which estimates the battery capacity of robots and the battery condition by the hierarchical fuzzy-neural network. As sub-project three focuses on developing an approach to explore the unknown environment by using many sensors, it also evolves functions of drawing special maps for group robots after obtaining the information in an unexplored environment, orientating and making the motion planning. Next, sub-project four will develop a non-contact infrared thermal imager method for the biometric identification, and moreover develop a disaster detection module that includes fire and posture detections after getting access to required accident information. Lastly, sub-project five is a research done in a group of campus patrolling robots by using infrared thermal imager techniques, which is a detection process that avoids unwanted consequences caused by factors such as nighttime or weak lighting. In other words, it can hence successfully detect the pedestrian, the position, velocity, and the orientation of the pedestrian in the campus. To conclude, the research group will combine all research findings of patrolling group robots; one believes that the accomplishment of this integrated project can be a big step in the development of robots in both the academic research field, as well as in industrial applications.