George M. Whitesides on Soft Robots and Soft Devices

okay why don't we get started so I want to welcome everyone to the Fulton schools of engineering deans Distinguished Lecture and we're gonna do something a little bit different perhaps in the sense that my role is to actually introduce the person who will introduce the speaker and the way to think about that is the Fulton schools were very large right we have about 350 faculty we're in the midst of another terrific hiring season and the power of a large faculty and having gotten great people from great places the people who've come over the past couple of years that people have been here for many years is the power of connections so we understand and we value partnerships and oftentimes that's with an agency Foundation but it's also with colleagues and many of our great faculty have come from great places themselves and in particular those connections then lead to an invitation for what I'm sure will be a very powerful lecture today so with that what I'd like to do is introduce one of our faculty professor Barbara Smith she's a faculty member in the School of Biological and health systems in health systems engineering and she was a member of dr. Whitesides lab so that gives a personal touch and a personal Flair that she can relate and she'll introduce our speaker so with that let me bring on professor Smith Barbara hi hi it is an absolute pleasure to be here today and an honor to introduce professor Whitesides I'm gonna start off in the beginning so but I'm gonna make it quick promise so professor Whitesides got his started as academic career at Harvard University and graduated in 1964 with a PhD from Caltech and for the next 19 years he's a professor at MIT and shifted over to Harvard since 1982 he's been a professor at Harvard he has over 1,100 publications get ready for the numbers he has mentored over 300 students both graduate students and postdocs and has invented over 50 is on over 50 different patents started over 12 companies and those companies have a net worth of more than 20 billion dollars he is extremely well known in the field he is part a member of the National Academy of Sciences and in 1998 professor white size was awarded the national merit of science so this is a man that we all know and we've all read his work and probably in soft lithography and micro fluidics point-of-care Diagnostics and today we're gonna hear about it in soft robotics and he's a man that started multiple fields and probably if you have it in your lab today its links back to professor Roy tides lab it's a great thinker and it's been a pleasure to know him for these last few years and with that I'm gonna pass it over to George Whitesides so join me in welcoming professor Thank You Deborah the Barbara left the lab how many years ago three I'm still exhausted those aren't even know it so those of you who are here for a while you have a exciting time ahead of you I want to talk about an entry into a particular kind of robotics and I know there are a lot of people who know a lot about robotics here and some who maybe know a little bit less about robotics so I'm going to try to make a mixture just a little bit at the beginning of introduction to why robotics is interesting to me and why our particular type of robotics is interesting and then a few results that are older which sort of give an idea of how rapidly the field is moving and then some stuff toward the end which is all new which concerns a key element of this area which is when you have a robotic system it's made of materials the materials do things typically and historically under the control of computers and sensors and things of this kind and one of the key areas here is can you simply replace the computers and sensors by the materials that is can you make the robot run itself based on the properties of the materials and why do you care and the answer is simpler cheaper more robust easier to work with so I'm a believer philosophically in the notion of simplicity and I'll come back to this as we go along all right so let me make a few remarks about soft robotics and actuators I want to start with a question which is my generic question about everything in science and engineering which is who cares and then spend some time on why soft and what is a soft robot and soft robot and actuator and you know what do you do with them and one of the things you do with them is you hold your glasses on your neck and I should say that the couple of a month ago I had the second my two cataract operation has done and I can't decide whether I see you more poorly with my glasses on or my glasses off but actually I think I'll leave them on for the while and see what happens so but if I keep squinting in odd ways it's not a neurological defect it's not a stroke about to happen not good it's just that and then what is material science have to offer so let me just start by making a remark about robotics in general when you hear people talking about robotics you're particularly in the popular press now there's a great deal of concern about the issue of robotics replacing people in work is this a good idea or a bad idea what happens when all work goes to machine this is this a good idea or a bad idea and I want to just give you a example of a historical precedent in this which is this and in the mid 30s there was a class of robots actually two classes of robots introduced which were washing machines and dishwashers which released about 50% of the population of the United States otherwise known as women from doing dish washing and washing machine like stuff to do teaching and being CEOs and running telephone systems and doing whatever people do but the point is not do robots displace people from jobs but do robots allow people to leave unpleasant jobs and to take on more pleasant jobs what's an example you need to make things real I look look out at you most of you with gratitude and you say well why and the answer is that I'm getting older the health care system has already spent down all the money that I put into it so there's nothing to keep me going my geriatrics decline except you so I expect you to be on the edge of your seat working hard for me and what does that have to do with this and the answer is that one of the big problems in Western Europe in the United States and Japan is going to be elder care and there's a lot of elder care Alzheimer's Parkinson's as just age which have to do with you know dealing with stuff that's actually not necessarily pleasant for everyone so let's replace a lot of that to the benefit of everyone with what machines can do and let people do something else the trick in this is what is the else that people are going to do so we desperately need to find new work for people to do and that's again for you all to do this is important okay what would be an example of another job that you would like to replace this is my idea of the end of human existence this is an Amazon packing house and it's a weird mixture of people who are doing a mind-numbingly boring job which is picking stuff out of bins where the bins are brought to them by robots that are mobile robots and it's really really intensely boring and what I want to do is to bring you up to date on the technology which actually it doesn't solve this problem yet but it's pointed at least in part toward the question of how do you take a bin of mixed stuff and then pick out the pieces that you want for different purposes which may seem like an interesting problem but it's not alright so why soft robots but if you talk to people who make robots robots in general and ask them what is their wish list what you want the first thing that everyone says is collaboration and what collaboration means in the trade jargon of robotics is you want robots that have the characteristic that they can work safely and cooperatively with human beings and if any of you have ever looked at the pictures that you see of a Toyota assembly line it is full of devices that are sort of half the size of the room roughly which control and auto body as it's being constructed in three dimensions with perfect dexterity with arms that by the way if they have them swing this way and you're in the way sure you often never notice this you're there they're absolutely not cooperative and so if you're talking about Alzheimer's patients it has to be something that is sort of like a human hand so that's what collaborative IDI means how do you how do you bring that to robotics and that you know it would seem possibly to imply softness and that's a part of the story a second thing that people want is simplification that is they want simplified controls they want systems to be adaptable they don't want to have complex elegant computers on their system they want to have nothing there ideally they want the system to do it by themselves the fewer parts there are the fewer parts there are to go wrong and that's an interesting part and for those of you spent a very interesting hour talking to people who love the complexity of computer-based control I admire people who love the complexity of computer-based control I want nothing to do with it if possible so that's fine so that brings up the idea of the computer as controller as material this controller and one of our ambitions in this is to make systems that have what we call reflexive materials I mean I there's the sometimes hear the phrase intelligent materials and I don't understand how a material can be intelligent but I do understand how a material given as a given circumstance can quickly and autonomously do something get shorter and get harder get faster again whatever it can be so how do you design that into materials and leads to some very interesting problems and then a bunch of other stuff like low-cost and lightweight and if you're doing just to take an example if you're doing surgery you do not want to wash your robot when you're finished because it's expensive and you don't always get it to work you'd like to throw it away well that requires a certain class of materials and so on you know there are a lot of fairly obvious things there the stimulus for all this stuff is from our point of view the scientific stimuli is first and foremost biomimetic that is we look at biology and we ask if we're gonna make something that's soft and behaves like a human what we do have to do to make it be soft and behave like human well we're a long way from that but that's the basic underlying idea we don't start with the idea of how you build an elaborate controls with the start with the question of how a pair of human hands Auctions and then ask if you can do something with that simplicity and what we call stackability stackability is a word that we've made up which actually I think it's an interesting word if you think about a computer system it starts with transistors a transistor individually is actually a pretty complicated device I mean it does very subtle things but they are now manufactured in such a fashion that you can get billions of them at a time with zero failures so you don't have to worry about transistors it's now a done deal so you now take transistors and using the miracles of a photo lithography you can make them into integrated circuits that haven't been in of them and that's probably one of the great constructions of the human mind works perfectly so you can forget about them you make microchips and you just buy the microchip and you take the microchips and you put those together and you make computers or whatever you have and they work so well you forget about them at each level you take a technology which can be very complicated but you make it work so perfectly that the next level of user never has to think about how it works and so the question with robotics to me is that if it is going to be something that is everywhere the individual pieces are going to have to work layer on layer perfectly and what that means is this few controllers as few computers as few everything of that sort as you can possibly have simplicity is really helpful in the world what else can I say it's another question of just saying it's a question of I have two other things being able to read because I can't both see you and read at the same time one is the question of tethering there are arguments about whether you want to robot that's entirely self powered it was we're self powered except we've got to eat every one soil so with computers and with my computer and with other systems you build batteries in ok I mean that that works but many robotic systems are actually run with plugs into the wall or inlets to power systems of other sorts mostly pneumatic so I'm not at this point very concerned with whether it's tethered or non tethered that's just a question put the power source the other issue is a word that's a little bit more complicated in its meaning which is autonomous and autonomy doesn't have a really clear meaning but in this context what it means is that the system faced with the necessity to do something does it on its own it's not controlled in doing it it just does it so it's autonomous it makes its own choice in a sense when faced with a task and I'll show you a lot of systems that do that so all interesting problems now I want to show you one of my favorite movies which of suiko runs which is in non fold robot but this is the picture that's really worthwhile looking at watch the way this moves and it shifts its center of gravity it rebalances it does everything that it does in a way that I think that if you blur your eyes you can't tell it's not alive I mean really looks like a horse or a pig or something like that doing you know pigs don't throw concrete blocks but you get the idea and it's breathtaking the clever of the thing you have to understand is that it has no brain there is only a collection of gyros and accelerometers in that and there's no central processor in no central intelligence so so what you say well the trick is take something like that which works so well that physically you can't really tell whether it's alive or not and combine it with artificial intelligence and then wait 10 to 20 years I don't really know what we're going to see but I don't think we're going to be able to tell those constructions from things that are alive so we're doing something that's really interesting which is we are doing what's called in the trade we're evolving a peer competitor that is a system in this case not alive or at least not at the moment which can do the same kinds of things that we can do in some cases better though maybe it looked different and maybe it'll have other inner character istakes but as soon as we start we and technology start doing that then it's an issue now let me just give you a little bit of a guide to some of the things that come in a bit this is what we would all say was a biomimetic system come on go backwards here biomimetic system that is a bi cept it's not relative to most of you who were young it's not an impressive bicep but it's a bicep to the question is can you mimic that and I want to show you a method of mimicking it but it's not trivial it's actually pretty interesting but it falls in this area of stackability it's based on nonlinearities of material in a quite subtle way but absolutely reproducible works perfectly and that's what it's going to look like so we're going to make something in which the operative device is that sort of purple thing which I'll show you how it works in a bit now one more last kind of background thing this is for those of you who are thinking about being entrepreneurs this is actually important we're interested in simplicity to design to make all the rest of those kinds of things but you also want to have a technology that has the characteristic that the end of the day somebody cares about it you know it's actually not produced just a paper but it's produced something that has an impact on society and how do you do that and I'm not in general in the business of shilling for books but let me show for a book because it's a very good one to read it's this thing that you see here by a guy named John Garnier and it's a history of Bell Labs and I hate histories of Bell Labs because they're typically physicists advertising for physics in a way that says you know brilliant physicists operating unconstrained by financial constraints invented solid state physics invented the transistor in vented the internet chase the way your children's brains are wired and all the rest of that and what the guy who is the head of the laboratory there bill Baker said is they were brilliant there's no question that Shannon and Bardeen and the rest these guys were fantastic however they were hired to and signed up to make a more profitable telephone system they did all of that work because people in New York wanted to place bets on hog futures in San Francisco and they couldn't do that so the issue there is that they were doing it not for the greater glory of science but to solve a problem and the problem was how do you make a system of switches that worked perfectly for transcontinental communication now what's the lesson for people who are interested in starting technology because there's no question that Bell was fantastically profitable at that and mr. Baker talked talks at length about innovation and all talks about innovation or much of anything else international conflict love and hate have four components and the reason for that is I have four fingers that I can hold up easily and a fifth is superfluous so what are the four components well the first one in his view is you need science and yes I think he felt you could get science almost anywhere it wasn't very expensive and very hard to get so you just get some science and then you do a step which he calls invention and what I think he meant by invention was largely making a functioning prototype of what you think you want to do so you've got to do the science and the science generates this prototype and then you do something that's hard and that's what he called development and that's engineering setting up a manufacturing line doing regulatory clearance thinking about all this stuff that you have to make when you have something that looks like a product so you get finished with all that expensive it takes time and so now you we've started with science and you have a product and his sharply focused question is what is the value now that you've gone through all this you've got your product what is the bad of that and he says zero where's the only thing that counts in his view is the market and what a market is is a social construct that has the interesting characteristic that you perform that most intimate of human interactions which is somebody writes a check for something because you have it and they want it and that is you know it's actually the market is sometimes misunderstood with people think about market and money as something that you get personally actually in almost all entrepreneurial stuff the market and the money that you get personally is such a small amount of money it isn't even worth working about worrying about what mostly happens in the market is that somebody says I like what you've got I'll pay you for it then you take that money and you make it even better and that means they'll buy more of them and they'll pay you for it more and there's a cycle in which the stuff gets better and better and better and if that doesn't happen you don't have anything so one of the ways in which you make a market is that you find something that's an entry level product that people want and they'll pay you for I'm going to tell you a little bit about that so let me just summarize some early work to get to a point which is what you see on the left-hand side here is a picture of a what we call a starfish gripper and I'll show you how this works in a bit it's an inflatable business basically some blow up balloons here's some raw egg you lower this raw this gripper on the raw egg and you inflate the balloons and the arms of the balloon and wrap around the egg and the nice thing about that is that you can't apply more pressure to the egg then you apply pressure to the balloons basically that's all you can do so you can't break the egg that actually it turns out to be a hard problem in robotics because you have to have sensors on the fingers to make sure you're not applying too much local pressure you have to have controllers you have to do a lot of other stuff it can be done but it's not trivial it's trivial to do this it's trivial so the part that's interesting is this we invented this stuff and published a paper on it in 2011 or something like that much to the resistance of the journal are the editors the reviewers and in 2016 the company that came out of all this this is a initial product and I'll show you what these things look like in a moment but for those of you who do this to go from a first publication to a commercial product solving a problem in six or seven years is really amazingly short for a manufactured thing and that worked only because we've stripped out all the computers all the controllers everything except something that's about as simple as you can make it in terms of material science works ok so here's how you make these things you you simply this is probably let's start over here you you take this think of this as a balloon here's the cross-section and if you blow up the balloon here what happens is it swells and nothing very interesting happens this is thicker and it doesn't go anywhere and this gets bulging here and you end up with a shape like that on the other hand if you put a little reinforcement in here that can't stretch because paper can bend but it can't stretch but this now the thicker part can swell and so what happens is this stays constant this bulges and you now have made a gripper because you've taken something from there and it's done that and this is a very simple idea but it actually pretty work it works pretty well and the details of how you make this are relatively trivial the one thing that's made a big difference is that this is all molded we use 3d printers for molding and the 3d printers are perfect for making the Masters for these things so that we can have an idea what a structure looks like at 8 o'clock in the morning and have a prototype working at noon and a key thing in all cycles of invention is to make failures occur as rapidly as you possibly can so a key thing is not to succeed fast but to fail fast if you do that you're in good shape so here is the device actually working drops down you blow air into it it goes down very gently wraps its way around and then picks up the the egg and it can't break it you just can't do that so it's a reflexive system if there's too much pressure at some point that pressure is adjusted by simply transferring pressure to someone else someplace else now there's a very interesting thing here which now starts to be something for the techies and the second part of the talk I'm going to show you a lot about this if you look at what's going on here is this particular thing blows up you'll notice something quite weird quite interesting and that is that these fingers which have uniform pressure from here to here start by curling at the end so it curls at the end why does it do that and that has to do with a situation that's crudely understandable by looking at this picture here's the you know one of these curl tentacles and the first bladder that's down here has nothing on one side and something on the other side so it's a little easier to blow up than the second bladder but there's a progression which means that this will blow up first and then that will propagate down this band it's not completely obvious but it's the way it works this is a non-linearity of materials and it's very useful and I'm going to spend a fair amount of time talking about other ways of doing that now how about reality let me show you a real system this is where these things are in a glamorous market which is handling celery and dough balls and things of that sort this is in real time and the only thing that we do is to make the effector on the end of this when I say we we mean the company but the characteristic of that system is that the grippers are compliant and they adjust themselves to the shape of the celery you don't have to know what the exact shape of the celery is all you have to do is get approximately there and it works so I mean it works most of the time but that's okay and your tomatoes tomatoes are an example of something which is soft and actually very easily damaged this does about the same thing that you do if you were to do it here only it does it faster and it doesn't get terribly bored while it's going on and getting in reality you don't always get all of them but there's another one of them just wood down station and then if you want to pick up cooked sausage links or you know other things of this sort well you know you sort of get the idea but the point of interest here is it does slippery things it does the slipperiest thing it turns out you can find is skin to chicken parts they're very slippery I suspect the fish are the same way and you know vegetables and dough balls and all kinds of things of that sort and so the market for this turns out to be food processing now is that a big exciting market yeah attractions enormous because the food industry wants to take people out of food processing lines and then the thing beyond that is what you do in CVS and things like that while you're unpacking crates so you're doing sweaters and bottles and all that kind of stuff these are potentially enormous businesses and it's not what I'm interested in and not what I do but it's nonetheless very interesting and it provides the market now you're what we're going to do to talk about the technical stuff I want to talk about buckling elastomeric beams and you may see the is that really interesting and the answer is yes it's really interesting negative pressure which means vacuum I showed you how to blow things up what happens if you suck them out and then nonlinearities and materials response these are sort of our three things for the moment and let me start with beams you all know that if you take a concrete post and you have a post then you start pushing on the post nothing happens for a while and you push hard enough built enough of a load eventually the post crumples and once it's crumpled once it's buckled and broken it's not fixable it's you know this is irreversible you see the same kind of thing here with damage to an autumn Oh I mean all this is was with steel beams to protect the inhabitants but once it's damaged it's damaged but this is a rubber beam and when you push on it it buckles but the buckling is reversible and this very simple idea that you go from something that's buckling with a rigid material where the entire community of material scientists thinks of it as a mechanism of failure you go to a soft material it becomes reversible and it becomes the mechanism of actuation so often good ideas are just sort of looking at something and say well everybody assumes it's that so it must be that it's not more complicated than that here is the sort of scientific basis of this let me show you what's going on there that pink thing that you see is a composite of air and a silicon polymer so it's holes in a piece of silicon polymer and our colleague and friend katya bartholdi studies this stuff and you notice as you squeeze it what happens is it does a pretty complicated thing these holes break up into two lattices of things this way in this way and there's a phase transition and so if you look at the stress and versus the strain in this region where it's just closing it does this and then closes over and does that so actually pretty interesting behavior for a structural polymer but it turns out this is complicated to replicate because you're trying to hold it with your hands and squeeze it uniformly on all sides so Dan yang who is working on this had a very good idea which was you take one of these systems of things like this you put a thin elastomeric film on the top and bottom and then you suck on it and what happens when you do that is that atmospheric pressure is uniform everywhere on the outside so automatically you apply completely uniform pressure on every exposed outside piece so it converts the problem of making this kind of motion reproducible into something that's trivial and when you do it you find something that's interesting now what I want you to do is to look at this to the extent you can see it here and think of these pieces here as beams you know they're holding things apart and as this as you suck it out what's going to happen the beams will collapse and when they do what they will do is rotate so this will be a way of converting a simple actuation which is the which is the yeah which is just applying vacuum to this system what it does is to convert that into something where you have rotary motion so it's a very complex conversion a very complex conversion of motors motions let's see if I can get that to do it I forget one no but this this will serve as well this serves illustrate the interesting nonlinearities with that vital task of picking up a toy elephant and what happens here is we have one source of actuation which is negative pressure here this is connected to vacuum this thing lowers on to the elephant please lower onto the elephant if you would be so good and when it does that what it does is the same motion takes the two arms and moves them in opposite directions so it's actually a lot of different stuff going on with a very simple input and that's what we like to see okay now to get back to my arm what we did here was to build a structure consisting of a series of beams that are rigid and then beams that are sort of more bendable when you suck on this these spaces decrease in size when that happens the whole thing crumples and when it crumples they get shorter so this is what goes on and let me show you with that piece which is the kind of thing that there we go so you suck on it goes up it's picking up a five kilogram weight so it's actually a pretty hefty piece so what that gives you is a linear actuator that enables us to do this now you say is this the crown of creation well I don't know if you look at it it turns out that that's probably the closest thing anyone has come to mimicking the function of muscle it's about the same efficiency the same speed the same kind of thing and it's very simple to make once you know how to make it is perfectly reproducible these systems will go through tens twenty million cycles with no evidence of degradation and performance at all which you can't get with most flexible systems requiring cleaning and lubrication so they're very very reproducible systems and here the motions this is not a very interesting motion from many points of view but it actually is pretty subtle to be able to do that so that's one example of all this the efficiency I'll pass by now let me go to this is all useful let me go now to one of our projects which we do for in principle amusement we call it life that never was and what does that mean it means that if you look at the creatures that are around you find that they walk they hop they swim they do whatever they do and that's all fine but you can imagine that there are lots of other ways they might get around I mean many kinds of motions why aren't any of those used and it's there's been no good way of answering that kind of question because you couldn't make any of them and what we are now able to do is to make all sorts of simple systems that might have occurred in some organism at some point and see whether it does anything interesting and what I'm going to do is to show you one test case in this which suggests that there's more than there than I would have thought it's actually pretty interesting so this is the this is work that's based on a collaboration we have with one of my colleagues whose name like many rock singers has one name his name is Mahadevan and you say anywhere in the world of condensed matter of physics you say the name Mahadevan and people say oh it's like you know all the great rock singers so so maja look and who's Maha we use Mahadevan because his second real Indian name is about that long it has 27 syllables and it's not pronounceable bajas very interesting guy and he got interested in the question of the following questions suppose I take a coffee cup and I put a sheet of plastic on the top of it and I put my finger here and I push what happens I try to shove this sheet down into the coffee cup the answer is that if you do it the right way it'll fold in a minute has to crumple up to do that and that that crumpling up is a buckling that's another kind of buckling so we got interested in this kind of buckling because one can easily add make imagine an organism that just made a sheet of stiff something collagen or whatever it might be and put some mechanism on for pushing on the middle so what we're going to do is to make that kind of system but what we're doing here is simply pulling on the background we're sucking on a bladder here and when we do that this thing undergoes this complex motion now those of you who are robotics people asked if you had a series of you know cables and stepping motors and things like that and you tried to replicate the complexity of that motion it would be very difficult to just look at the front leg it is actually pretty good you say all right so it crumples so what and fair question technology probably not so important these will go through many cycles here's the thing that's important and that is that when you take this crumpling sheet and you simply put some actually that actually gallops pretty good there and the question is if it's so easy if all you need is one sheet some primitive legs and one device for applying pressure why isn't somebody used just actually one of the better walkers we have so I don't know I mean I think it looks pretty good and here's the bottom of this and you just even does a sort of pop so we will now go off and see what happens and you can make things that look like that there are unusual shapes and you can make it out of things of this kind you know the it isn't restricted here's a sort of an interesting thing as well and that is if you take one of these sheets and you bend it then when it does this crumpling it will always crumple there if you bend it in two places it will crumple in both you bend it in three places it'll crumple in all three so what you can do is to program this not numerically by just bending it because it remembers the stresses that you put into it we haven't yet determined whether you can unprogrammed and reprogram it in some other way but this is I think a really interesting idea of how you do programming in a purely analog way without any of the numeric stuff that goes into what we normally call programming and it again works quite smoothly and the let's see if I've got a movie here yeah that's a system that's been programmed to get four four buckles and it will happily go on and give that kind of shape if you ask me what do I want to do with that shape okay uh and here was the picture this is a little bit out of order before I showed you the thing with legs I should really have showed you this and imagine that you're an organism and you're on a sandy beach and what you really would like to do is when there's a shadow of the past predator and what's the simplest way that you can imagine to go somewhere else the answer is probably this is as good a way as anything this moves along pretty well it doesn't have much except one sort of pushing edge and one sliding edge that could be fixed up a little bit and one actuation which is easily done by even very primitive organisms and I would think that if I were an organism I would be very happy being able to move that far rather than just sitting there and waiting and saying come eat me and so you know the extension is that you go on from there and you put on legs but that was not all right here I've seen that now what next what's next in this area I've given you a sort of an idea of where stage zero is rapid commercialization stage one which is the discovery that buckling is interesting stage two is the idea that there may be things that you can do is buckling and there are many different kinds of buckling the people have never thought of because they'd never thought of buckling as being something that anyone would want to do you normally want to prevent it from being done so where are we now we're gonna exploit these kinds of things but the major thing to point out is that if you think about what I've shown you it's all elastomers of this sort they're like rubber bands PDMS is is commonly used in breast replacement and tissue replacement it's that polymer so if you have a feeling for those objects you sort of know what the consistency is but here's something which has the characteristic that it's another elastomer and it's very soft now relative to these things and then I happen to be particularly interested in this which is the basis of what we're calling hypothetically hard soft robots if you're looking for something that is really really tough you don't use steel you use rubber particle reinforced rubber and if you look the things that form the basis for truck tires for heavy-duty mining and things like that are all particle reinforced rubbers and so that's things we want to work on a very interesting number in this is you say we apply pressure we apply vacuum how much how much is required 7 psi how much is 7 psi 7 psi is half an atmosphere so it's half of what we live in it's getting to be what you see towards the top of a very high mountain and you say well that's not very much is it and there's an interesting number which is the difference in pressure between a 747 in flight the difference in pressure between the top of the wing and the bottom of the wing is about 7 PS is the amount that we use and we say that can't be the case because you're lifting up this entire airplane but remember pressure is force times area pressure times area and if you simply do a envelope back-of-the-envelope calculation and ask how many square inches there are on the wetted wings of a 747 and multiply it by 7 it turns out to be in the order of a million pounds which is about 1/2 747 wings however however it's really easy to go to 100 psi and if you do at a different kind of way you can go to a thousand psi and you know these things are a couple of hundred psi and we're all quite comfortable using bicycle tires Racing bicycle tires that are 500 PS 122 goes bang it's a big bang but nobody gets hurt so 100 psi at the scale that we're using you know something like this is 100 pounds of force and something like this gets to be a ton and so I think it's going to be actually pretty straightforward to build systems of this sort which will do one of our group's ambitions which is to tear an automobile in half I hope we do it soon because they don't have that long to wait but I really would like to see that happen because I can use that to confound some of our enemies so but if you look at this there's an entire program here going from very you know sort of ordinary elastomers to very soft to very tough and very hard which will make classes of these soft robots of almost every sort that you can think about whoo interesting now just very briefly for those of you who are students what do you care about market opportunities ecommerce thinking things agriculture all above all sorts food of all sorts consumer products and then for heavy industries if we start doing that some of these other things lifting working with toxic chemicals we're taking in radioactive environments places where people don't want to work this these will work well there and actually in many of those you can't send computer systems in if you tried to send computer system into into Fukushima you fried the electronics essentially instantly here you don't care because these are really easy things and they crawl over stuff biomedicine is interesting and hazardous jobs of a variety of sorts so a very wide range of things that you can do with soft systems some of the more intellectual things I think this idea of controlling replacing controls by the properties of materials it's actually a good idea now I don't know how far it's going to extend but I have the idea that taking computers out of stuff rather than making computers more complicated is actually a good thing to do simply because it increases robustness and decreases price and also computer systems you know they typically don't work very well under water and in very hot environments and things of that sort nonlinear middle era teas and materials a problem that's a very interesting one if you talk to the conventional hard robotics guys about this they say well this is all fine but you're using compressed air and they say that with condescension dripping from their lips see well what should I use you say you should use electricity have you heard of it and you know all this conversation is not kindly intended you say well why do you want to use electricity and they say you didn't hear me I said electricity I said no I hear you say you said electricity why electricity you said you're not listening I said electricity so this gets to be not a very useful conversation but I don't know why one necessarily wants to use electricity I actually like pneumatics because they're not much more complicate they're safe they don't short when you put them underwater there are a lot of things you can do with them so I like that and then a few other things that are that you can read and finally there are a whole bunch of scientific stuff which you can read while I'm talking but one of the nice things about this comes at the bottom here these are so easy to make these systems are so easy to make that you can actually think about doing stuff like this and we have interns come every summer and your interns come expecting to wash dishes and instead we have them make their own soft rocks off robot but it's so easy to do they love it they just think this is a great thing to do and I think when you make science both interesting simultaneously interesting and really easy for people to work with it's a good thing to do for the SMT Enterprise and so that's one of the nice things about this no I've already shown you my prejudices there let me finish with this this is a list of people who have been involved in this project over the course of now probably a decade and you won't know the names but what I would like you to look at is not even what they did but where they come from and you'll find some Americans there but a lot of Canadians and Iranians and you know all kinds of people German turkey turkey Canada Iran Canada Scotland whatever and the argument is an argument is very familiar here but I think we really as a community need to beat on this and that is these are smart people who are curious who are doing the work that's made this possible because they're smart people and they're curious there aren't an infinite number of them in the world and the United States needs to compete for them so anything that can be done to block this issue of making it hard for foreign immigrants particularly people who are particularly talented to get it into the United States we need to really resist we can put up with a lot of other stuff but basically science progresses on the basis of people and this project to me we could never have done it without people who came with totally different backgrounds in totally different cultures so they did the work there are some people who reluctantly paid and I'm reluctantly grateful it works out well and I thank them and I very much thank you and I love talking to engineers it gives me a feeling that reality is not entirely lost in the world so thank you for the invitation to be here so thanks very much for a really terrific seminar we do have some time for questions and I think one idea sticks the microphone into people's hands if they have questions who wants to go first yeah and your part buckling sheets or you're buckling sheets how do you get a directional motion why don't they just stand in place or wander around randomly of course the buckling of course makes the circuit makes a circular sheet and symmetrical and so it will always buckle in the same place and if we needed buckles in the same place what happens with these things is the tailing edge digs in so it can't go backwards and the front end because it is less in contact can go forward so that's you know this sort of basic notion these are also pretty good swimmers you can make jellyfish out of them and things of that kind but they're intrinsically asymmetric as soon as they buckle fascinating talk and so you have so many ideas so where are the sources of your infinite number of ideas and also boundless energy and we have a lot of students here and what made you to do your PhD and two questions okay so so you can encourage our students right there there will be a late dinner at midnight as soon as I finish this okay know the origin of this project is actually a little bit interesting I've worked for years with various groups in the DoD and one of the groups I've worked with is something called DARPA which many of you know the Defense Advanced projects research agency and they've been interested of course the military's been interested in robots of all sorts for a long time but there are problems that come up which current systems won't actually handle and an example of one is how do you get a robot that can walk through surf onto a beach crabs do it just fine but robots don't do it at all well so the question was if you look at existing life-forms and you say it doesn't have to look like me it doesn't have to be bipedal it doesn't have to have two arms and two legs but you're free to look at anything that's alive and ask what's there then since I like simplicity my view is you look at starfish you look at worms you look at snails and in particular you look at octa points because there's nothing more sophisticated than an octopus or cuttlefish and so we set out after convincing after some years of discussion we got a DARPA project whose idea was to I won't give it it's informal name but anyway the idea was you want to start with an object that looked like a coke can would spontaneously unfold itself into a flat sheet spontaneously walk across the floor find a door jam crawl under the door jam and then reconstitute itself onto a coke can and the other side and you can see why darkman might be interested in that it was not particularly interested in that but that's okay but we set out as in many good projects set out to do it and we didn't get the whole thing of course to work but every component of it worked that is we were able to individually get all the components to work and so once we sort of had this idea and you know a lot of the ideas are just trivial but for example are you going to use chemical actuation are you going to use external pressure are you going to use electricity we're going to use mass how are you going to do it once you sort of settle on some of these basic ideas which can take a while then a lot of the rest of it actually almost doesn't require any thought and for those of you who are students the thing to remember is that if you're the first in some area then everyone who does works in the area afterwards works for you if you're the second in the area you always work for whoever is first so there's a lot to be said for being first and does that answer your question I mean intrinsically I'm immensely lazy so you know the idea is to do things that are first so you don't have to worry about all the people who are second and third and fourth and fifth so trying to replace any human activity by a robot and robot doesn't have a brain doesn't have a heart so in cases where you know you know human beings and the people you know don't like to do certain things definitely robot is helping yeah risky job you know but in some cases a service like go to restaurant all the robots are you know serving with food is not so there is no human touch yep so what is your thought on that well my first thought is I know many humans who have neither a heart nor a brain so I don't see this necessarily as a disadvantage but you know it turns out to increasingly their cultures in which people are actually more comfortable operating with machines than operating with with with humans I mean apparently it's it's one of the big motivations well in a clear example this emoji you know where emoji came from Mogi was invented in Japan because the Japanese have a difficulty in expressing emotion in text and emoji provided a way of expressing a kind of emotion in text without having to write something out and of course it's spread worldwide because it does all kinds of interesting things but different cultures have different feelings about this and so I would in fact far prefer to talk with a robo caller than I would with the human where they feel no sense of guilt and saying clunk with a robo collar and with a human there's always you know the fact that here's this poor person who is being paid in calls per hour and should I listen politely for five minutes so they get credit for the call or should I not so I mean I think the point is legitimate but the issue of whether whether you're ten years down the road and the the the simulator of a person who's invariably friendly and invariably Kurtis and recognises immediately what kind of joke that you want and you know can be tipped on your ubirr account so you don't have to worry about anything so system I don't know what we're going to come out liking best but I think it's actually a it's a question that could make you a little uneasy I plan to be dead by then so it doesn't make me uneasy at all this come here a little bit of philosophy question so in terms of the general biomimetics idea do you believe that in some day everything of organisms or lives can be decoded by physics some of my best friends are physicists but physics doesn't solve all problems so I don't know the answer to that I think the the issue that's interesting here is not actually the human part the robot part the machine part the IR part it's when you put them together what happens when you actually do a fusion of living and nonliving and artificial intelligence and all the rest of that what does that look like and I don't think we can look at that with any more clarity of foresight then you would have had if you could some of you will remember Quinn transistors came in little cans with three legs and you soldered the three legs into the device they would I had absolutely no idea what would come of transistors and the trick in this as with as with transistors is going to be to make all of this stuff these functions so reliable so simple so inexpensive that you can use any number of them that you want in any way that you can reconfigure them and see what comes out of it and we can't predict that so what you want in a good area is something where there's a big there in principle that I'm positive of I don't know what form it's going to take so that's up to you to figure out and many of the other fields you talked about there are ways of modeling those systems like in transistors we use like HDL and arrow log to model that for these new soft systems how is there any modeling that exists currently the answer is very primitive katia does some of this stuff with regular arrays but I think that one of the interesting characteristics of these very nonlinear systems is that it may be that we don't have good analytical models but we do it by AI that is you simply you know make the thing let it's you know let it wander all of its over all of its phase space give it every possible motion let the system remember all of that and then you never know what it's doing maybe but it does it now downstream that sounds bad because you're you know you're trying to stack a system of things that are that are basically being operated on principles that you don't understand and you have to think about how one layer interacts with another I don't know what to do about that that might be a problem or not so right now nonlinear mechanics and elastomers is a pretty primitive area because nobody's been interested enough to do it that doesn't necessarily mean it can't be done I have listened to you many times and I always get excited by the new vision you're painting and it is always thoroughly enjoyable I had a question which you might have answered partly with your response in the last one as we move from simple so to speak functions I like those you described into more complicated ones where certain complicated logic maybe involved what sort of scientific technical directions in the areas of materials should we somehow start pursuing asking questions in them in order to provide code and code logic more complicated logic to simple mechanical motions I think there are a couple of answers to that very good question one of them is that I think we'll be able to build systems which are logical we have three or four now in which we can undergo all of the you know end or whatever kinds of simple logical operations whether those those are not going to be useful in competition with electronic systems for doing logic but they may be useful for doing simple information processing of some kind in building the machines but I think more interesting is I visit us I emphasize the idea that these I would like to take sensors out of these but that's not going to work you know you're going to have to have sensors of some kind so building materials that have the characteristic that they build in new types of sensor information so I mean just suppose when I strain something it becomes fluorescent and I can look at the degree of strain by the fluorescence or I know I can build these systems using exactly the same technology to have lasers built in to have wave guides built in to have wires built in hence head magnetic fields built in micro fluidic system is built-in and one of the nice things about the soft systems is that for example we can build a tentacle and the tentacle couldn't go out and sort of sniff this pile of white powder and then can sample it and then can drip sulfuric acid on it if that's the right thing to do that's not trivial to do when you're dealing with a system it's hard because you don't know what you're going to be dealing with so I think the whole notion of building multifunctional materials that are nonetheless the last American will stand up to this is actually pretty short term and very interesting and will dramatically expand the field so perhaps one more question George M just out of curiosity if we think of our brains will you put our brains in your cat they were it's a soft soft material soft matter but he has very complicated function and he learns and he does things by itself that's why would you say about a brain um I have a sort of theory that I sort of believe we're just the following that if I think about I'm as a chemist so I start thinking about molecules and if I know think about a cell a cell is a collection of chemicals and the chemicals are reacting the chemicals themselves the molecules are not alive the reactions are not alive but the collection of molecules and reactions is alive we don't really know what that means but it's it's to me very interesting it may just be something that happens you know an analogy is that life is just like a flame you're a complex set of reactions stable so long as you feed it oxygen methane or glucose and you know peanut butter and it creates life and that's what life is and then maybe that's all that sort of life is if that's true because that's true which I actually think it is then what is a brain a brain is a collection of cells and the cells are just collections of reactions that are happening so us sentience self-awareness love hate all of that may just be a spontaneous activity of complex systems that happens to have had a extended persistence in in the environment in which we live it's a thought that makes me as uneasy as it probably makes others of you but you know is there a difference between you know a flickering flame of wax gas in the in the swamp in Siberia and Beethoven thinking about the 9th well I though that's something to for those of you who are young yeah think about it so on that note it's good to have a homework assignment of sorts one more thing before we conclude a there's going to be a reception right outside so as you come down the stairs and you exit please join us for the reception we also have a momento Arizona State University IRA Fulton schools of engineering and appreciation of George M Whitesides deans lecture series April 4th 2018 let's thank dr. Whiteside it's one more time [Music] [Applause]

Channel: Ira A. Fulton Schools of Engineering

Views: 997

Rating: 4.7777777 out of 5

Keywords:

Id: szuEE6EdrXg

Length: 65min 30sec (3930 seconds)

Published: Fri Apr 20 2018