Learning Sequential Acquisition Policies for Robot-Assisted Feeding
Learning Sequential Acquisition Policies for Robot-Assisted Feeding
复制标题
DOI：
10.48550/arxiv.2309.05197
复制
发表时间：
2023-09
期刊：
影响因子：
--
通讯作者：
Priya Sundaresan;Jiajun Wu;Dorsa Sadigh
中科院分区：
文献类型：
--
作者：
Priya Sundaresan;Jiajun Wu;Dorsa Sadigh
一个提供进餐帮助的机器人必须使用各种餐具进行专门的操作，以便拾取和喂食一系列食物。除了这些灵巧的低级技能，辅助机器人还必须在很长的时间内按顺序规划这些策略，以清理盘子并完成一顿饭。之前的机器人辅助喂食方法引入了高度专业化的食物处理基元，但无法将它们组合在一起。与此同时，现有的方法，长期的操作缺乏灵活性，高度专业化的原语嵌入到他们的框架。我们提出了视觉行动规划OveR序列（VAPORS），一个框架，长期的食物收购。VAPORS通过在模拟中利用学到的潜在板块动力学来学习高级行动选择的策略。为了在真实的世界中执行顺序计划，VAPORS将动作执行委托给可视化参数化原语。我们验证了我们的方法在复杂的现实世界中的收购试验涉及面条收购和双手舀果冻豆。在38个盘子中，VAPORS比基线更有效地获得，在实际的盘子变化中（如浇头和酱汁）进行了概括，并在对49个人进行的调查中定性地吸引了用户的喂养偏好。代码、数据集、视频和补充材料可以在我们的网站上找到：www.example.com。
A robot providing mealtime assistance must perform specialized maneuvers with various utensils in order to pick up and feed a range of food items. Beyond these dexterous low-level skills, an assistive robot must also plan these strategies in sequence over a long horizon to clear a plate and complete a meal. Previous methods in robot-assisted feeding introduce highly specialized primitives for food handling without a means to compose them together. Meanwhile, existing approaches to long-horizon manipulation lack the flexibility to embed highly specialized primitives into their frameworks. We propose Visual Action Planning OveR Sequences (VAPORS), a framework for long-horizon food acquisition. VAPORS learns a policy for high-level action selection by leveraging learned latent plate dynamics in simulation. To carry out sequential plans in the real world, VAPORS delegates action execution to visually parameterized primitives. We validate our approach on complex real-world acquisition trials involving noodle acquisition and bimanual scooping of jelly beans. Across 38 plates, VAPORS acquires much more efficiently than baselines, generalizes across realistic plate variations such as toppings and sauces, and qualitatively appeals to user feeding preferences in a survey conducted across 49 individuals. Code, datasets, videos, and supplementary materials can be found on our website: https://sites.google.com/view/vaporsbot.