基於搜尋演算法之規劃移動式機器手臂尋找與夾取遮蔽目標物

隨著機器人與相關研究的發展由工廠漸漸走向辦公室及家庭環境，機器人在人類生活環境中服務與互動顯得十分重要，機器人應該要能做為人類生活中的幫手並提供各式服務，而在各項服務中，搜尋並遞送物品對使用者來說是很實用的一項技能。為了達成此功能，前人的研究多假設物品放在開放空間中，機器人在室內環境中以影像搜尋目標物並抓取之，過程中機器人規劃出最佳的路徑移動並且徹底觀察環境中的每個角落。然而，在真實世界中，物品很可能被其他物品所阻擋以致單純的影像搜尋不可能找到物品，因此，近期的文獻提及了以機器手臂移除障礙物以達成目標物搜尋之規劃，這些研究成果規劃機器人移動障礙物的順序並探索後方空間。在本篇論文中，我們結合了以機器手臂移除障礙物與視覺主動搜尋之規劃，提出了一整合兩種方式之物品搜尋系統，機器人可以變換位置觀察環境，亦可以使用手臂移動障礙物來進行搜尋。本篇論文提出之規劃概念是以最小化預期搜尋時間以及在發現目標物後最小化目標物抓取時間為目標，使用A*演算法並在搜尋空間內進行取樣以達成在有限時間內完成規劃。此外，本論文加入了視覺回授以確保機器人準確地執行計畫。最後，我們透過在書架內物品阻擋的環境中搜尋目標物的實驗來驗證本論文提出方法之優越性。

關鍵字

機器手臂操作；搜尋物體規劃；辦公室機器人

並列摘要

As robots and robotic researches marching from factory to office and home, the ability of robot to interact with complex human-living environment becomes pivotal. To show its value, the robot should be able to do various tasks as an assistant in human-living environment. Searching and delivering object in indoor environment is one of the tasks practical to user. Previous study mainly focused on visual search of objects in indoor environment. The search is performed by a mobile robot which plans a best route to observe the environment and discover the target object. However, in real world, objects may be occluded by other objects or structures, which means pure visual search is impossible to find these targets. As a result, some recent works discussed the object search method by removing objects that block and hide the target object. In this thesis, we propose an object search planning system that combines visual and arm manipulation search. The robot can either reposition one of the accessible object with its arm or move its platform to view the environment from a different position to discover the target object. The concept of planning is A* Planning which minimizes the expected time to discover the target and then the time to grasp the target in clutter after its discovery. Visual sensor feedback is included to assure the accuracy of each action performed by the robot. We evaluate the proposed approach with experiment in the scenario of object search in a shelf environment where objects may occlude or block access to one another.

並列關鍵字

robot arm manipulation ； object search planning ； office robot

參考文獻

[1] Y. Ye and J. K. Tsotsos, "Sensor Planning for 3D Object Search," CVIU, vol. 73, pp. 145-168, 1999.

[5] J. Ma, T. H. Chung, and J. Burdick, "A probabilistic framework for object search with 6-DOF pose estimation," Int. J. Rob. Res., vol. 30, pp. 1209-1228, 2011.

[11] P. E. Hart, N. J. Nilsson, and B. Raphael, "A Formal Basis for the Heuristic Determination of Minimum Cost Paths," IEEE Transactions on Systems Science and Cybernetics, vol. 4, pp. 100-107, 1968.

[12] M. A. Fischler and R. C. Bolles, "Random sample consensus: a paradigm for model fitting with applications to image analysis and automated cartography," Commun. ACM, vol. 24, pp. 381-395, 1981.

[14] I. Sipiran and B. Bustos, "Harris 3D: a robust extension of the Harris operator for interest point detection on 3D meshes," The Visual Computer, vol. 27, pp. 963-976, 2011.

國際替代計量

基於搜尋演算法之規劃移動式機器手臂尋找與夾取遮蔽目標物

全文下載

主題瀏覽