請用此 Handle URI 來引用此文件:
http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/4766
標題: | 基於同時調整多策略以最佳化零力矩點軌跡實現人型機器人之行走平衡 Future ZMP Trajectory Optimization with Simultaneous Multi-Strategy Adjustment for Humanoid Walking Balance |
作者: | Bang-Cheng Wang 王邦丞 |
指導教授: | 王傑智(Chieh-Chih Wang) |
關鍵字: | 人型機器人,走路,零力矩點, Humanoid,Walking,ZMP, |
出版年 : | 2015 |
學位: | 碩士 |
摘要: | 人形機器人走路時一件富有挑戰性的控制問題。現今研究已發現可以基於調整控制輸入、步伐大小和步伐長度三種策略的未來零力矩點(Zero-Moment Point)軌跡來增進機器人平衡。然而,我們發現現今方法由於未考慮多策略之間的相互關係,因此增進平衡的能力是受限制的。在此論文中,我們分析多策略之間的相互關係,並且提出基於同時調整多策略以最佳化零力矩點軌跡實現人型機器人之行走平衡,而最佳化未來零力矩點軌跡是透過最佳化能量函數所產生的。最後,本研究與現今頂尖的零力矩點軌跡調整方法比較,並證明此演算法能夠讓機器人在更艱難條件下維持平衡。 Humanoid walking balance has been a challenging issue in the control field since the balance can only be achieved by considering the dynamics of the system. It is found that the modification of the future ZMP trajectory based on the multiple strategies, which are the adjustment of the control input, the step size and the step duration, can enhance the robot balance. However, the capacity of the disturbance compensation is limited due to the fact that the correlation among the strategies is not considered. In this work, the correlation among the three strategies is analyzed, and the future ZMP trajectory optimization with simultaneous multi-strategy adjustment for humanoid walking balance is proposed. After defining an energy function for the robot balance, an optimized trajectory based on the simultaneous adjustment is generated. By comparing our work with the state-of-the-art future ZMP trajectory modification approaches, it is shown that our approach can maintain the robot balance under more severe situations. |
URI: | http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/4766 |
全文授權: | 同意授權(全球公開) |
顯示於系所單位: | 資訊工程學系 |
文件中的檔案:
檔案 | 大小 | 格式 | |
---|---|---|---|
ntu-104-1.pdf | 743.95 kB | Adobe PDF | 檢視/開啟 |
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。