Localizing Complex Terrain for Quadruped Robots using Adaptive Submodularity;Localizing Complex Terrain for Quadruped Robots using Adaptive Submodularity

Please use this identifier to cite or link to this item: http://ir.lib.ncu.edu.tw/handle/987654321/93513

Title:	Localizing Complex Terrain for Quadruped Robots using Adaptive Submodularity;Localizing Complex Terrain for Quadruped Robots using Adaptive Submodularity
Authors:	張軒旗;Chang, Hsuan-Chi
Contributors:	數學系
Keywords:	次模性;多足運動;最大覆蓋問題;觸摸定位;Submodularity;Legged Locomotion;Maximal coverage problem;Touch Localization
Date:	2023-04-19
Issue Date:	2024-09-19 17:09:52 (UTC+8)
Publisher:	國立中央大學
Abstract:	近年以來，多足機器人的運動學已被廣泛研究。該研究領域的主要問題在於環境的不確定性。為了解決這個問題，本研究提出了一種稱為Adaptive Submodularity with Hypothesis Pruning(ASHP)的方法，將不規則地形上的運動問題重新定義為覆蓋Perlin Noise領域中的問題。多足機器人能夠在沒有外部感受器的情況下，漸漸地適應在復雜地形上的運動。Adaptive Submodularity在本文中被應用於預測地形的樣貌，並提供相關的理論保證值。模擬和實驗顯示出本文提出的方法相比於Random Selection有更小的預測誤差，且相較於其他的模型(RMA)有更高的成功率及穩定性。;The locomotion of the legged robot has been widely researched in recent years. The main issue in this research area is environmental uncertainty. To overcome this issue, this research proposed a method called Adaptive Submodularity with Hypothesis Pruning(ASHP), which reformulates the locomotion problem on irregular terrains as the coverage problem in the Perlin domain. The legged robot is able to adaptively select locomotion over complex terrains without exteroceptive sensors. The adaptive submodularity is utilized to predict the terrain with theoretical guarantees. The simulations and experiments demonstrate that the proposed approach has less prediction error and a higher success rate than the benchmark, the experiments also show that ASHP is more robust than the benchmark method(RMA).
Appears in Collections:	[Graduate Institute of Mathematics] Electronic Thesis & Dissertation

Files in This Item:

File	Description	Size	Format
index.html		0Kb	HTML	14	View/Open

社群 sharing

Loading...