So, robots. Robots can be programmed to do the same task millions of times with minimal error, something very difficult for us, right? And it can be very impressive to watch them at work. Look at them. I could watch them for hours. No? What is less impressive is that if you take these robots out of the factories, where the environments are not perfectly known and measured like here, to do even a simple task which doesn't require much precision, this is what can happen. I mean, opening a door, you don't require much precision.
说到机器人。 机器人可被编程进行 数百万次的重复任务,而几乎不出错, 这对我们来说很难做到,对吧? 工作中的机器人很迷人。 看看它们。 我可以盯着看几个小时。 没有这种感觉吗? 不那么迷人的是, 如果这些机器人离开工厂, 进入一个不同的环境,不像 在工厂这样进行过精确的了解和测量, 去完成一项不需要什么 精度的简单任务, 结果可能是这样的—— 我是说,就开个门, 不用考虑什么精度。
(Laughter)
(笑声)
Or a small error in the measurements, he misses the valve, and that's it --
或者测量数据里有个小误差, 它错过了阀门,就这样了——
(Laughter)
(笑声)
with no way of recovering, most of the time.
多数情况下,根本无法恢复。
So why is that? Well, for many years, robots have been designed to emphasize speed and precision, and this translates into a very specific architecture. If you take a robot arm, it's a very well-defined set of rigid links and motors, what we call actuators, they move the links about the joints. In this robotic structure, you have to perfectly measure your environment, so what is around, and you have to perfectly program every movement of the robot joints, because a small error can generate a very large fault, so you can damage something or you can get your robot damaged if something is harder.
这是为什么呢? 多年来, 机器人的设计一直 在强调速度和精度, 这已经演变成了 一种非常特定的构架。 如果观察一个机器臂, 它是一组精准设计的刚性连杆, 还有马达,也叫致动器, 马达让连杆绕着关节移动。 在这种机器人结构中, 你必须完美地测量工作环境, 了解周围环境里有什么, 还要对机器关节的每一个运动 完美地编程, 因为一个小误差就会 产生非常大的故障, 会损坏别的东西,或者 如果别的东西比机器人硬, 就会损坏你的机器人。
So let's talk about them a moment. And don't think about the brains of these robots or how carefully we program them, but rather look at their bodies. There is obviously something wrong with it, because what makes a robot precise and strong also makes them ridiculously dangerous and ineffective in the real world, because their body cannot deform or better adjust to the interaction with the real world. So think about the opposite approach, being softer than anything else around you. Well, maybe you think that you're not really able to do anything if you're soft, probably. Well, nature teaches us the opposite. For example, at the bottom of the ocean, under thousands of pounds of hydrostatic pressure, a completely soft animal can move and interact with a much stiffer object than him. He walks by carrying around this coconut shell thanks to the flexibility of his tentacles, which serve as both his feet and hands. And apparently, an octopus can also open a jar. It's pretty impressive, right?
那么我们来说说这些机器人。 不要去想这些机器人的大脑, 也不要想我们对它们的 编程有多仔细, 仅仅观察一下它们的身体。 其中显然存在一些问题, 因为使机器人精确并强壮的东西, 同时也让它们在现实世界中 变得极度危险且低效, 因为它们的身体不能变形, 也不能更好地适应 与真实环境的交互。 那么,不妨设想一种 截然相反的方式, 让它比周围的所有东西都软。 也许你觉得,如果它很柔软 就什么都干不了, 这似乎有点道理。 但大自然的说法则相反。 例如,在海底, 数千磅的静水压下, 一只周身柔软的动物 可以移动比它硬得多的物体, 并与之相互作用。 它带着这个椰子壳到处走, 就因为触角是柔软的, 触角既是脚也是手。 显然,章鱼还能打开罐子。 很神奇,对吧?
But clearly, this is not enabled just by the brain of this animal, but also by his body, and it's a clear example, maybe the clearest example, of embodied intelligence, which is a kind of intelligence that all living organisms have. We all have that. Our body, its shape, material and structure, plays a fundamental role during a physical task, because we can conform to our environment so we can succeed in a large variety of situations without much planning or calculations ahead.
但是,这样的行为不仅仅是由 这种动物的大脑实现的, 也是由它的身体实现的, 这是一个明显的例子, 也许是最明显的例子, 体现了嵌入式智能, 这是所有生物都有的一种智能。 我们也都有。 我们的身体,它的 形状、质地和结构, 是躯体活动的基础, 我们能主动适应环境, 所以能在各种各样的 状况下完成任务, 而不必事先做很多计划或计算。
So why don't we put some of this embodied intelligence into our robotic machines, to release them from relying on excessive work on computation and sensing? Well, to do that, we can follow the strategy of nature, because with evolution, she's done a pretty good job in designing machines for environment interaction. And it's easy to notice that nature uses soft material frequently and stiff material sparingly. And this is what is done in this new field of robotics, which is called "soft robotics," in which the main objective is not to make super-precise machines, because we've already got them, but to make robots able to face unexpected situations in the real world, so able to go out there. And what makes a robot soft is first of all its compliant body, which is made of materials or structures that can undergo very large deformations, so no more rigid links, and secondly, to move them, we use what we call distributed actuation, so we have to control continuously the shape of this very deformable body, which has the effect of having a lot of links and joints, but we don't have any stiff structure at all.
那么为什么不把某些嵌入式智能 应用到机器人中, 让它们摆脱对计算和感知的 过度依赖呢? 要实现这一点, 我们可以遵循自然法则, 因为大自然已经通过进化 设计出了非常好的 与环境相互作用的机体。 很容易注意到,大自然 经常使用柔软的材料, 而很少使用坚硬的材料。 这就是这个新领域, 或机器人领域所实现的, 叫做“柔性机器人学”, 它的主要目的不是制造超精密机器, 因为我们已经有这种机器了, 而是要让机器人能够面对 真实环境中的意外情况, 从而能够进入真实环境。 机器人之所以柔软, 首先是因为它的柔性躯干, 它由可以承受很大变形的 材料或结构制成, 不再用刚性的连杆, 其次,为了让它们移动, 我们使用分布式驱动, 我们必须持续控制这个 非常易变形的躯干的形状, 它的效果就像 有很多个连杆和关节, 但却根本没有任何刚性结构。
So you can imagine that building a soft robot is a very different process than stiff robotics, where you have links, gears, screws that you must combine in a very defined way. In soft robots, you just build your actuator from scratch most of the time, but you shape your flexible material to the form that responds to a certain input. For example, here, you can just deform a structure doing a fairly complex shape if you think about doing the same with rigid links and joints, and here, what you use is just one input, such as air pressure.
可以想象,造一个 柔性机器人的过程 与构造刚性机器人非常不同, 后者需要用特定方法 去组装连杆、齿轮和螺钉。 而对于软体机器人, 大多数情况下, 只需要从头构建致动器, 将柔性材料塑造成 对特定输入做出响应的形式。 例如让这个结构发生形变, 如果想用刚性连杆和关节, 需要相当复杂的形状去实现, 而这里使用的只是一个输入值, 比如大气压。
OK, but let's see some cool examples of soft robots. Here is a little cute guy developed at Harvard University, and he walks thanks to waves of pressure applied along its body, and thanks to the flexibility, he can also sneak under a low bridge, keep walking, and then keep walking a little bit different afterwards. And it's a very preliminary prototype, but they also built a more robust version with power on board that can actually be sent out in the world and face real-world interactions like a car passing it over it ... and keep working.
好的,我们来看几个 很酷的柔性机器人吧。 这个是哈佛大学研发的 一个可爱的小家伙, 他靠着沿身体施加的压力波走路, 由于是柔性的, 他可以从矮桥下溜过, 再继续走, 然后稍微变换姿势继续走。 这是最初的原始模型, 他们还构建了更强大的版本, 搭载了动力, 该版本可以与外部真实的环境的互动, 比如一辆车在它身上碾过... 还能继续工作呢。
It's cute.
太可爱了。
(Laughter)
(笑声)
Or a robotic fish, which swims like a real fish does in water simply because it has a soft tail with distributed actuation using still air pressure. That was from MIT, and of course, we have a robotic octopus. This was actually one of the first projects developed in this new field of soft robots. Here, you see the artificial tentacle, but they actually built an entire machine with several tentacles they could just throw in the water, and you see that it can kind of go around and do submarine exploration in a different way than rigid robots would do. But this is very important for delicate environments, such as coral reefs.
还有机器鱼,它能像 真鱼一样在水中游动, 因为它有一个柔软的尾巴, 同样利用了气压进行分布式驱动。 那是麻省理工学院的作品, 当然,还有机器章鱼。 它实际上是柔性机器人学 这一新领域开发的首批项目之一。 你看到的是人造触须, 其实整个机器都是用几根触须打造的, 把机器扔进水里, 可以看到它到处走, 探索海底世界, 它的行为方式与硬机器人很不一样。 这对于珊瑚礁等脆弱的环境非常重要。
Let's go back to the ground. Here, you see the view from a growing robot developed by my colleagues in Stanford. You see the camera fixed on top. And this robot is particular, because using air pressure, it grows from the tip, while the rest of the body stays in firm contact with the environment. And this is inspired by plants, not animals, which grows via the material in a similar manner so it can face a pretty large variety of situations.
让我们回到陆地。 你看到的这个画面是 我的斯坦福同事开发的 正在成长的机器人。 相机固定在上面。 这个机器人很特别, 因为利用气压,它从顶端生长, 而身体的其他部分 与环境保持紧密接触。 这是受到植物的启发,不是动物, 它们以相似的方式 传过环境中的障碍不断生长, 因此可以面对多种多样的情况。
But I'm a biomedical engineer, and perhaps the application I like the most is in the medical field, and it's very difficult to imagine a closer interaction with the human body than actually going inside the body, for example, to perform a minimally invasive procedure. And here, robots can be very helpful with the surgeon, because they must enter the body using small holes and straight instruments, and these instruments must interact with very delicate structures in a very uncertain environment, and this must be done safely. Also bringing the camera inside the body, so bringing the eyes of the surgeon inside the surgical field can be very challenging if you use a rigid stick, like a classic endoscope.
我是一名生物医学工程师, 也许我最热衷的应用 是在医学领域, 我能想到的与人体近距离互动方式 应该就是实际进入身体内部了, 例如微创手术。 在这方面,机器人对 外科医生的帮助很大, 因为他们要进入人体 实施手术,就必须 利用很小的开口和平直的器械, 这些器械必须与 非常精细的结构相互作用, 在非常不确定的环境中, 并且必须保证安全。 另外,要把相机带入体内, 让医生看到手术区域内部, 如果用硬的器械, 比如传统的内窥镜, 是非常有挑战性的。
With my previous research group in Europe, we developed this soft camera robot for surgery, which is very different from a classic endoscope, which can move thanks to the flexibility of the module that can bend in every direction and also elongate. And this was actually used by surgeons to see what they were doing with other instruments from different points of view, without caring that much about what was touched around. And here you see the soft robot in action, and it just goes inside. This is a body simulator, not a real human body. It goes around. You have a light, because usually, you don't have too many lights inside your body.
我曾与欧洲的研究小组一起, 开发了这种手术用 自摄像柔性机器人, 它与传统内窥镜非常不同, 能利用组件的柔性来移动, 该组件可以向各个方向 弯曲,也可以伸长。 外科医生实际上已经 在用它从不同视角 观察自己对其它器械的操作, 而不用太担心碰到周围的东西。 这里可以看到操作中的柔性机器人, 它就这么进去了。 这是人体模拟器, 不是真正的人体。 它可以四处移动, 还能提供光照,因为通常 身体里通常没什么光亮。
We hope.
最好别有。
(Laughter)
(笑声)
But sometimes, a surgical procedure can even be done using a single needle, and in Stanford now, we are working on a very flexible needle, kind of a very tiny soft robot which is mechanically designed to use the interaction with the tissues and steer around inside a solid organ. This makes it possible to reach many different targets, such as tumors, deep inside a solid organ by using one single insertion point. And you can even steer around the structure that you want to avoid on the way to the target.
有时,外科手术甚至 只用一根针就能完成, 目前我们正在斯坦福研究 一种非常灵活的针, 是一种非常微小的柔性机器人, 它的机械结构被设计成能够 利用与组织的相互作用, 在实体脏器内来回移动。 这让我们有可能到达 实体脏器内部深处的 许多不同组织,例如肿瘤, 而且只需要使用单个切入点。 你甚至可以在通往目标组织的路线上 绕过想要避开的结构。
So clearly, this is a pretty exciting time for robotics. We have robots that have to deal with soft structures, so this poses new and very challenging questions for the robotics community, and indeed, we are just starting to learn how to control, how to put sensors on these very flexible structures. But of course, we are not even close to what nature figured out in millions of years of evolution.
显然,对于机器人技术来说, 这是一个非常激动人心的时刻。 我们有了柔性结构的机器人, 这对机器人界提出了 新的、非常有挑战的问题, 事实上,我们刚刚开始 学习如何控制, 如何在这些非常灵活的 结构上放置传感器。 当然,对于大自然在数百万年的 进化中创造的奇迹, 我们的发现根本不值一提。
But one thing I know for sure: robots will be softer and safer, and they will be out there helping people. Thank you.
但有一点我可以肯定: 机器人将变得更加柔软和安全, 它们将在现实生活中 为人类提供帮助。 谢谢。
(Applause)
(掌声)