This website requires JavaScript.

Demonstration-guided Optimal Control for Long-term Non-prehensile Planar Manipulation

Teng XueHakan GirginTeguh Santoso LembonoSylvain Calinon
Dec 2022
摘要
Long-term non-prehensile planar manipulation is a challenging task for robotplanning and feedback control. It is characterized by underactuation, hybridcontrol, and contact uncertainty. One main difficulty is to determine contactpoints and directions, which involves joint logic and geometrical reasoning inthe modes of the dynamics model. To tackle this issue, we propose ademonstration-guided hierarchical optimization framework to achieve offlinetask and motion planning (TAMP). Our work extends the formulation of thedynamics model of the pusher-slider system to include separation mode with faceswitching cases, and solves a warm-started TAMP problem by exploiting humandemonstrations. We show that our approach can cope well with the local minimaproblems currently present in the state-of-the-art solvers and determine avalid solution to the task. We validate our results in simulation anddemonstrate its applicability on a pusher-slider system with real Franka Emikarobot in the presence of external disturbances.
展开全部
图表提取

暂无人提供速读十问回答

论文十问由沈向洋博士提出,鼓励大家带着这十个问题去阅读论文,用有用的信息构建认知模型。写出自己的十问回答,还有机会在当前页面展示哦。

Q1论文试图解决什么问题?
Q2这是否是一个新的问题?
Q3这篇文章要验证一个什么科学假设?
0
被引用
笔记
问答