CRII: RI: Enabling Manipulation of Object Collections via Self-Supervised Robot Learning

虽然在杂乱的现实世界环境中操纵单个对象已经受到了极大的关注，但直接操纵对象集合的问题尚未得到探索。该项目研究机器人可以在多大程度上自主操纵这些对象集合。该项目促进了适用于家庭机器人助手的自主操作方法。这种辅助机器人将对提高老年人和患有某些退行性疾病的人的生活质量产生重大影响。此外，该项目中研究的机器人技能适用于在自然或人为灾害中受损的地区进行操作，需要清除建筑物瓦砾和其他碎片。该项目支持为低收入学校的儿童开发和提供互动机器人讲座，教授他们计算机编程的基础知识。该项目的研究目标是使机器人能够连续操纵和推理物体组。该项目的假设是，将对象集合视为单个实体，可以实现数据高效的、自我监督的接触位置学习，以推动和抓取分组对象。本项目研究了一种用于自监督操作学习的新型神经网络架构。卷积神经网络模型将传感数据和机器人手配置作为输入。网络学习预测给定输入的操作质量分数作为输出。当呈现一个新的场景时，机器人可以通过评估当前的感觉数据来执行操作推理，同时直接优化网络在不同手部配置上预测的操作分数。该项目支持制定实验协议和收集相关数据以供传播，以促进在操纵物体收集方面的研究活动。

While manipulation of individual objects in cluttered, real-world settings has received substantial attention, the problem of directly manipulating collections of objects has been left unexplored. This project investigates to what extent robots can autonomously manipulate such object collections. This project facilitates autonomous manipulation methods suitable for use in home robotic assistants. Such assistive robots stand to make a substantial impact in increasing the quality of life of older adults and persons with certain degenerative diseases. Additionally, the robot skills investigated in this project are suitable for manipulation in areas damaged in natural or man-made disasters, where building rubble and other debris need to be cleared. The project supports the development and presentation of an interactive robotics lecture for low-income school children, teaching them fundamentals of computer programming.The research goal of this project is to enable robots to manipulate and reason about groups of objects en masse. The hypothesis of this project is that treating object collections as single entities enables data-efficient, self-supervised learning of contact locations for pushing and grasping grouped objects. This project investigates a novel neural network architecture for self-supervised manipulation learning. The convolutional neural network model takes as input sensory data and a robot hand configuration. The network learns to predict as output a manipulation quality score for the given inputs. When presented with a novel scene, the robot can perform manipulation inference by evaluating the current sensory data, while directly optimizing the manipulation score predicted by the network over different hand configurations. The project supports the development of experimental protocols and the collection of associated data for dissemination to stimulate research activity in manipulation of object collections.