查詢結果分析
來源資料
相關文獻
- Real-time Trajectory Planning for Robot Menipulators a Spherical Modeling Approach Integrating Neural Netowrks and Artificial Potential Fields
- 用於實驗階段之兩足機器人定位電路設計與實現
- 採摘柑橘果實機器手臂之研製
- 撞球機械手臂之研製
- 漫談機器手臂應用技巧
- 漫談以機器手臂為主架構之自動化系統
- 3D雷射視覺應用於汽車工業機器手臂之電弧銲接
- 機器手臂應用在番茄大小選別過程中搬移作業之研究
- 三維機器手臂避碰路徑規畫之電腦輔助設計與模擬
- 視覺系統引導之機器手臂採果之研究
頁籤選單縮合
題 名 | Real-time Trajectory Planning for Robot Menipulators a Spherical Modeling Approach Integrating Neural Netowrks and Artificial Potential Fields=以整合神經網路及位能場的球狀模擬法作實時機器手臂軌跡規劃 |
---|---|
作 者 | 馬恆; | 書刊名 | 萬能學報 |
卷 期 | 19 1997.07[民86.07] |
頁 次 | 頁337+339-356 |
分類號 | 448.94 |
關鍵詞 | 整合神經網路; 球狀模擬法; 機器手臂; |
語 文 | 英文(English) |
英文摘要 | This paper describes a new spherical modeling method and a modified potential field method for dealing with geometric complexity and kinematic redundancy in the robotics trajectory planning problem. The spherical modeling method employs a neural network mapping to describe the boundary information of a complex object as a collection of spheres that fill the object volume. The potential field method, as employed here, takes the objects thus defined and identifies a safe path of least potential. Feasible robot trajectories are determined by the Newtonian motion law under the influence of the potentials defined. A model including a PUMA 560 robot and three designated obstacles with increasing geometric complexity were created in Autocad R12 to test the spherical modeling and potential field methods proposed. Simulation results showed consistent efficiency for varied degrees of geometric complexity in collision detection, i.e. around two orders of magnitude faster for the case of an object with a contour surface than the nearest benchmark alternative in Autocad, and demonstrated the capability to determine feasible trajectories for the six-degree-of-freedom robot. The system was implemented on a real PUMA 560 which then successfully avoided cylindrical obstacles to achieve a specified goal position and orientation. |
本系統中英文摘要資訊取自各篇刊載內容。