RI: Small: Optical Skin For Robots: Tactile Sensing and Whole Body Vision

这个项目将使机器人能够感觉到它们所触摸的东西。关键的想法是在机器人体内安装摄像头，在机器人皮肤变形时向外观察，并通过机器人皮肤观察附近被抓住或避开的物体。这种方法解决了几个挑战：1)使用数百万像素实现接近人类的分辨率（一百万个生物传感器），2)减少抓取和操作过程中的遮挡，并在撞击前检测障碍物，以及3)保护昂贵的电子设备和线路，同时允许更换磨损或损坏的廉价皮肤。人类每个月都会更换皮肤外层。有一种理论认为，目前的机器人之所以如此笨拙，是因为它们的皮肤几乎没有感觉。触摸时能感知的机器人将成为更好的仆人，更有用，尤其是在照顾老年人和其他需要帮助走路、穿衣、清洁或喂食的人时。在家庭和工作场所的许多任务中都需要柔软的触感。如果机器人能感觉到自己在做什么，教机器人完成新任务和机器人学习将会容易得多。如果机器人能感觉到意外接触或它们施加的力的大小，它们也会更安全，更安全。这种方法的可能应用包括感知家具和汽车内饰，感知衣服，感知工具，感知地板，墙壁和天花板。光学皮肤传感是一种实用、经济、可制造和可维护的触摸和帮助人们的传感方法。该项目的技术目标包括首先在机器人上建立并安装一个由大约100个现成的小型摄像机（小于1立方厘米）组成的网络，该网络能够收集信息，决定需要关注哪些视频流，并处理视频流以估计力、滑动和物体形状。传感系统中的布线、焊点和其他连接不必反复变形或面临循环或可变应力——这是故障的主要来源。变形或受到压力的材料（皮肤的外层）可以优化用于传感，抓取和操作，以及机械稳健性，并且在磨损或损坏时可以轻松廉价地更换。研究人员将通过共同开发一种手部“光学皮肤”和一种协同柔软的手来评估他们的工作。一个革命性的想法是在整个机器人身体上积极地分布高分辨率成像。这减少了遮挡，这是操作感知中的一个主要问题。考虑到成像传感器的低成本，不再需要将光学传感限制在红外测距仪（单像素深度相机）、线相机或低分辨率区域相机上。

This project will enable robots to feel what they touch. The key ideais to put cameras inside the body of the robot, looking outward at therobot skin as it deforms, and also through the robot skin to see nearby objects as they are grasped or avoided. This approach addresses several challenges: 1) achieving close to human resolution (a million biological sensors) using millions of pixels, 2) reducing occlusion during grasping and manipulation, and detecting obstacles before impact, and 3) protecting expensive electronics and wiring while allowing replacement of worn out or damaged inexpensive skin. Humans replace the outer layer of our skin every month. One theory as to why current robots are so clumsy is that they have little or no feeling in their skin. Robots that can feel when they touch will make better servants and be more useful, especially when taking care of older adults and others needing help walking, dressing, cleaning, or feeding. A soft touch is needed in many tasks in the home and workplace. Teaching robots to do new tasks, and robot learning,will be much easier if robots can feel what they are doing. Robots will also be safer, and safer to work with, if they can feel accidental contact or the size of forces they are applying. Possible applications of this approach include aware furniture and car interiors that feel what a human is doing or wants to do, awareclothes, aware tools, and aware floors, walls, and ceilings. Optical skin-based sensing is a practical, affordable, manufacturable, and maintainable approach to sensing for touching and helping people.Technical goals for the project include first building and then installing on a robot a network of about 100 off-the-shelf small cameras (less than 1 cubic centimeter) that is capable of collecting information, deciding what video streams to pay attention to, and processing the video streams to estimate forces, slip, and object shape. The wiring, solder joints, and other connections in the sensing system don't have to repeatedly deform or face cyclic or variable stresses--major sources of failure. The materials that do deform or are stressed (the outer layer of the skin) can be optimized for sensing, grasping and manipulation, and mechanical robustness, and can be easily and cheaply replaced when worn or damaged. The researchers will evaluate their work by co-developing an "optical skin" for hands and a synergistic soft hand. A transformative idea is to aggressively distribute high resolution imaging over the entire robot body. This reduces occlusion, a major issue in perception for manipulation. Given the low cost of imaging sensors, there is no longer a need to restrict optical sensing to infrared range finders (single pixel depthcameras), line cameras, or low resolution area cameras.