WEBVTT 00:00:03.497 --> 00:00:10.517 This is the Convergent Science Network podcast. Leading researchers in the domain 00:00:10.517 --> 00:00:16.797 of neuroscience, brain theory and technology are interviewed by Paul Verschure and Tony Prescott. 00:00:18.997 --> 00:00:27.957 This is Tony Prescott for the BCBT 2011 Summer School. and I'm talking to Federico 00:00:27.957 --> 00:00:32.997 Carpi from University of Pisa Research Centre and Rico Piaggio. Hello. 00:00:33.257 --> 00:00:38.317 So Federico, this morning you gave us a really interesting talk about the next 00:00:38.317 --> 00:00:43.397 generation or what might be the next generation of actuators for robots and 00:00:43.397 --> 00:00:44.317 many other applications. 00:00:44.877 --> 00:00:49.597 So before you tell me what these new actuators are like, can you explain why 00:00:49.597 --> 00:00:50.717 do we need new actuators? 00:00:50.777 --> 00:00:54.437 What's wrong with the tiny little motors that we've got already? 00:00:55.017 --> 00:00:59.957 Yeah, there are a lot of, let's say, problems that must be overcome with traditional 00:00:59.957 --> 00:01:01.077 actuation technologies. 00:01:02.277 --> 00:01:08.977 One of them certainly deals with the fact that they use stiff material, mostly metal. 00:01:10.057 --> 00:01:16.497 And so this means that they are using some constitutive material that are actually 00:01:16.497 --> 00:01:22.677 very, very far from the natural tissue that our bodies are made of. 00:01:22.677 --> 00:01:31.317 So we are talking about some stiff components as compared to soft tissue that makes our bodies. 00:01:32.397 --> 00:01:35.717 And this is particularly important because, and this aspect I would say has 00:01:35.717 --> 00:01:37.577 been completely neglected so far. 00:01:37.577 --> 00:01:46.617 Attention has been made only on the psychological aspect on one hand and the 00:01:46.617 --> 00:01:54.637 neurophysiological aspect of replacing human-like machines. 00:01:54.977 --> 00:01:59.337 But the body itself is really, really important. 00:02:00.517 --> 00:02:07.517 And another aspect, for instance, concern more technical issues like acoustic 00:02:07.517 --> 00:02:11.037 noise that traditional motors generate, 00:02:12.117 --> 00:02:18.457 which is not something really pleasant for users, 00:02:18.777 --> 00:02:25.377 especially when we consider technologies or systems that have to work in close 00:02:25.377 --> 00:02:28.577 contact with humans, for instance, rehabilitation 00:02:28.957 --> 00:02:31.297 system, like orthotics or prosthetics. 00:02:31.637 --> 00:02:37.457 And there are also other issues, for instance, power consumption, energy consumption. 00:02:38.657 --> 00:02:43.677 Basically, electric motors are really demanding in terms of energy consumption, 00:02:43.957 --> 00:02:50.577 which make this aspect challenging if we consider portable application, 00:02:50.857 --> 00:02:54.277 for instance, in terms of autonomy of the robot. 00:02:55.322 --> 00:02:58.502 So these and other issues 00:02:58.502 --> 00:03:07.622 suggest that we should look for some alternative technologies based on materials 00:03:07.622 --> 00:03:10.542 that are inherently capable of 00:03:10.542 --> 00:03:16.642 exhibiting any kind of useful response to a suitable electrical stimulus. 00:03:16.642 --> 00:03:22.602 In particular, we are looking for materials that are able to deform or change 00:03:22.602 --> 00:03:26.082 shape according to an electrical stimulus. 00:03:26.362 --> 00:03:31.042 So we are looking for so-called electromechanically active materials. 00:03:31.662 --> 00:03:38.222 So these would be in some way more similar to sort of animal actuators or muscles. 00:03:38.522 --> 00:03:44.022 Exactly. What are the nice properties of animal muscles that you think we want to try and copy? 00:03:44.022 --> 00:03:49.542 For sure, we are looking for materials that should be able to exhibit both strain 00:03:49.542 --> 00:03:54.602 and stress, active strain and active stress capabilities, as I said, 00:03:54.602 --> 00:03:57.042 in response to an electrical stimulation. 00:03:58.222 --> 00:04:03.862 And for sure, we are also looking for a combination of actuation and sensing. 00:04:04.022 --> 00:04:07.662 These two features should be integrated in the same material. 00:04:07.662 --> 00:04:16.602 So, in this respect, these smart materials should be able to self-sense their 00:04:16.602 --> 00:04:21.322 own state, mechanical state, in terms of, for instance, deformation. 00:04:22.442 --> 00:04:29.882 So we are opening here a different paradigm with respect to the state of the art. 00:04:29.982 --> 00:04:36.342 Nowadays we use actuators and sensors typically as different components in a robot, 00:04:36.502 --> 00:04:44.422 and the The two components should, of course, interplay and are part of the same system, 00:04:44.542 --> 00:04:50.322 but in any case they are different devices working with different principles 00:04:50.322 --> 00:04:52.622 of operation and frequently, 00:04:53.162 --> 00:04:58.122 they are also arranged in different positions, at different locations of the robot. 00:04:58.122 --> 00:05:02.522 On the other hand, in the human being, or in general in the animals, 00:05:02.782 --> 00:05:11.582 the device, let's call it like this, the device itself is both an actuator and a sensor. Our muscle, 00:05:12.498 --> 00:05:16.118 inherently have the sensing capabilities so it's 00:05:16.118 --> 00:05:19.258 the material it's the tissue itself that work both 00:05:19.258 --> 00:05:25.078 as an actuator and as a sensor so the so that our muscles are made up of of 00:05:25.078 --> 00:05:29.618 cells and each of those is in itself very complicated oh yeah so how can we 00:05:29.618 --> 00:05:34.058 use materials yeah to build something with well the idea of course is not to 00:05:34.058 --> 00:05:38.918 replicate the cellular structure yeah of the of the natural tissue, 00:05:39.478 --> 00:05:48.818 but the idea is to replicate the functional properties, so in terms of actuation and sensing. 00:05:50.198 --> 00:05:58.318 From this respect, the so-called electromechanically active polymers seems to 00:05:58.318 --> 00:06:03.018 be really, really interesting, and there's a lot of competence in Europe on that. 00:06:03.018 --> 00:06:07.898 So for somebody who's not a chemist, what's an electrally active polymer? Yeah. 00:06:08.058 --> 00:06:19.198 So they are just a piece of matter, a polymer, can be of different types of synthetic form. 00:06:19.658 --> 00:06:26.018 One of the most useful nowadays consists of so-called dielectric elastomers. 00:06:26.018 --> 00:06:33.018 Strictly speaking just a piece of insulating rubber nothing more like a piece of silicon. 00:06:33.938 --> 00:06:37.438 Which is an electrical insulator which can 00:06:37.438 --> 00:06:42.758 be deformed as soon as it is charged properly from electrical point of view 00:06:42.758 --> 00:06:49.818 and we can enter if you wish the physical principle i will try to to explain 00:06:49.818 --> 00:06:53.858 it in simple words basically you could explain very briefly yeah passing current 00:06:53.858 --> 00:06:55.278 through a piece of rubber can make a change. 00:06:55.278 --> 00:07:01.198 Well, actually, it is not really a matter of passing current because the principle 00:07:01.198 --> 00:07:03.338 is just an electrostatic effect. 00:07:03.658 --> 00:07:08.938 So in simple words, you have a layer of an insulating elastomer, 00:07:09.018 --> 00:07:14.618 as I said, a piece of silicon, for instance, and you cover this layer on the 00:07:14.618 --> 00:07:16.698 main surfaces with two electrodes. 00:07:17.578 --> 00:07:21.258 And these electrodes should be compliant, so deformable. 00:07:22.356 --> 00:07:26.216 And in this way, you have a capacitor, what the engineer called a capacitor, 00:07:26.496 --> 00:07:27.496 an electrical capacitor. 00:07:27.896 --> 00:07:34.896 As soon as you charge this capacitor electrically, the charges on the electrodes 00:07:34.896 --> 00:07:41.676 basically interact according to a simple electrostatic effect, so Colombian forces. 00:07:42.016 --> 00:07:48.016 So basically the plus and the minus attract each other and the charges of the 00:07:48.016 --> 00:07:51.236 same sign repel on the electrodes. 00:07:51.236 --> 00:07:58.636 This effect, combined together, squeezes the layer, the insulating layer, 00:07:58.896 --> 00:08:00.416 between the two electrodes. 00:08:01.276 --> 00:08:05.636 So we have, practically speaking, really a compression of the material. 00:08:06.076 --> 00:08:13.796 And since these elastomers have a constant volume, the compression along the 00:08:13.796 --> 00:08:19.856 thickness is parallel to a surface expansion, because the volume should be constant. 00:08:19.856 --> 00:08:24.756 In this way, we achieve a significant actuation, so a significant deformation 00:08:24.756 --> 00:08:28.896 of this capacitor as soon as it is charged. 00:08:29.156 --> 00:08:34.056 So the principle of operation is extremely simple, a purely electrostatic effect, 00:08:34.176 --> 00:08:37.596 which is known as Maxwell stress effect. 00:08:38.956 --> 00:08:43.096 It is not a piezoelectric effect, I want to remark. It is different. 00:08:43.476 --> 00:08:48.316 So the principle is that we're using this electric field to squeeze a piece of rubber. 00:08:48.616 --> 00:08:52.636 Exactly. Okay, and then how can we use that as an actuator to move things? 00:08:52.636 --> 00:08:54.336 Yeah, so this is the physical principle. 00:08:54.536 --> 00:09:00.456 Then we can build an actuator by changing, by opportunely designing the shape, 00:09:01.256 --> 00:09:03.516 what we call the configuration of the device. 00:09:04.456 --> 00:09:10.656 For instance, a simple piece of rubber with planar shape is a very elementary 00:09:10.656 --> 00:09:15.196 actuator, but it is useful for many tasks. 00:09:15.636 --> 00:09:21.416 Then if we for instance, if we stack in a sort of pile multiple layers, 00:09:21.636 --> 00:09:24.196 one on the top of each other, or the other, 00:09:24.856 --> 00:09:30.276 we achieve actually a pile, a stack, and this is very useful, for instance, 00:09:30.356 --> 00:09:36.556 to have larger deformations because each layer will contract and the whole pile, 00:09:36.556 --> 00:09:42.596 the whole stack of course will contract accordingly but you achieve a higher absolute, 00:09:43.276 --> 00:09:49.836 contraction of course because simply because you have a taller a bigger pile. 00:09:50.876 --> 00:09:56.436 That's it this is what we call stacked actuators for instance but there there 00:09:56.436 --> 00:10:03.016 is really a plenty of different configuration available nowadays many many groups 00:10:03.016 --> 00:10:06.216 have demonstrated a lot of them, including our group in Pisa. 00:10:06.376 --> 00:10:13.096 And for instance, these include a linear actuator or membrane-like actuator, 00:10:13.256 --> 00:10:18.276 bubble-like actuator, really, really any kind of configuration. 00:10:18.676 --> 00:10:23.656 The limitations are mostly in our imagination, I would say. This technology, 00:10:23.876 --> 00:10:24.876 this is really interesting. 00:10:25.016 --> 00:10:28.396 An interesting point, this technology is highly versatile. 00:10:29.603 --> 00:10:38.463 Scalable, and is really suited to be shaped according to the need of the specific application. 00:10:39.023 --> 00:10:43.723 I can see how you can get this thing to deform and therefore actuate. 00:10:43.843 --> 00:10:46.403 But how can it be intrinsically sensing? 00:10:46.843 --> 00:10:51.603 Yeah, you can use their sensing capabilities in two respects. 00:10:52.503 --> 00:10:58.023 One is the so-called piezocapacitive effect. So basically, as we said, 00:10:58.163 --> 00:11:00.583 the device is a deformable capacitor. 00:11:01.543 --> 00:11:07.023 So if you read the capacitance, the electrical capacitance of the device, 00:11:07.223 --> 00:11:12.643 while it is undergoing the deformation, you have a signal which is correlated 00:11:12.643 --> 00:11:14.023 to the deformation itself. 00:11:14.203 --> 00:11:19.883 So by reading the variation of the capacitance, you have an information on the 00:11:19.883 --> 00:11:26.563 deformation state of your device continuously. So the actuator is both an actuator, 00:11:26.643 --> 00:11:28.743 but also a piezocapacitive sensor. 00:11:29.743 --> 00:11:36.263 You could also exploit another concept, another principle, which is the piezoresistive effect. 00:11:36.483 --> 00:11:41.763 So in this case, you need to read a resistance, a variable resistance, 00:11:41.943 --> 00:11:49.483 like any conventional strain gauge that is used everywhere in industrial products. 00:11:49.483 --> 00:11:52.163 In this case as i said you have 00:11:52.163 --> 00:11:55.203 to read a resistance which can be either the 00:11:55.203 --> 00:11:58.763 resistance of one of the two electrodes or of 00:11:58.763 --> 00:12:02.063 the resistance of an additional layer conductive layer 00:12:02.063 --> 00:12:07.643 that you integrate in your device it is very very easy very easy to to achieve 00:12:07.643 --> 00:12:13.643 a sensing scene and this is extremely stimulating because as i said in this 00:12:13.643 --> 00:12:19.343 way we achieve an integrated sensor and actuator all in one. 00:12:19.963 --> 00:12:28.263 And this is from a functional point of view very similar to what is available in our natural tissue. 00:12:29.414 --> 00:12:34.654 So we can build these quite compact devices that can actuate and sense. Definitely. 00:12:35.054 --> 00:12:39.554 So what's the drawback? Because why aren't we putting them into all our machines now? 00:12:40.114 --> 00:12:44.954 There's still some problems to solve. Sure. As usual, there are advantages and disadvantages. 00:12:46.594 --> 00:12:50.494 For instance, you said properly that you can achieve very compact structure. 00:12:50.794 --> 00:12:54.754 That's why the first product, which has been launched this year, 00:12:55.354 --> 00:13:03.134 with a huge expected commercial impact is a product for consumer electronics. 00:13:03.494 --> 00:13:08.714 Maybe I cannot say here the specific company because can I say my line? 00:13:08.734 --> 00:13:10.314 I think you can. We can edit it out. 00:13:10.754 --> 00:13:19.454 Yes. But by the way, it is a big company which sells consumer electronics like cell phone. 00:13:20.094 --> 00:13:26.534 And in this case, This company has replaced the electrical motors that are used 00:13:26.534 --> 00:13:32.334 to provide vibration to the mobile unit with these material, 00:13:32.434 --> 00:13:36.394 electromechanically active polymers, in particular dielectric elastomer actuators. 00:13:36.534 --> 00:13:41.834 And they have produced these actuators as thin films, very, very thin. 00:13:41.934 --> 00:13:48.454 So this is very appealing in order to integrate them in a thin, 00:13:48.634 --> 00:13:51.354 portable device like a cell phone. 00:13:52.271 --> 00:13:56.991 These devices are very thin, very compact, they are very lightweight, 00:13:57.291 --> 00:14:03.331 and they consume very low power because they are capacitors, 00:14:03.371 --> 00:14:08.091 electrical capacitors, so they do not need to be driven with high current. 00:14:08.351 --> 00:14:15.511 That's a great advantage in order to save energy, so in order to enable a really 00:14:15.511 --> 00:14:18.351 portable application with low power consumption. 00:14:18.911 --> 00:14:24.311 So to get a vibration out of this sheet? What are you doing to the current in 00:14:24.311 --> 00:14:25.511 order to generate the vibration? 00:14:26.211 --> 00:14:30.671 Oh, yes. You have a sinusoidal voltage, for instance. 00:14:30.991 --> 00:14:35.371 So you charge and discharge your device. So it's very straightforward from the 00:14:35.371 --> 00:14:39.171 current that you supply to the behavior that you get. Sure, sure. 00:14:39.371 --> 00:14:43.111 There's a specific relation which is not linear, but it is specific. 00:14:43.111 --> 00:14:46.051 It is a quadratic response. response i mean the 00:14:46.051 --> 00:14:48.871 the stress that you can generate in this 00:14:48.871 --> 00:14:51.831 material is proportional to the 00:14:51.831 --> 00:14:54.971 square of the electric field so the electric 00:14:54.971 --> 00:14:58.571 pulses that we use when we move our muscles can can 00:14:58.571 --> 00:15:03.351 we use those similar kinds of pulses to drive these sorts of materials yes we 00:15:03.351 --> 00:15:08.831 can but it really depends uh on the application if it is useful or not so for 00:15:08.831 --> 00:15:11.811 instance we have a medical application for yeah for instance we have demonstrated 00:15:11.811 --> 00:15:16.731 but it is a very simple experiment that you can drive this material with electromyography. 00:15:17.331 --> 00:15:22.311 That's very easy, like any other actuator. So that's not a critical point today. 00:15:22.331 --> 00:15:26.591 So this would be where you would record signals directly off somebody's muscles. 00:15:26.771 --> 00:15:30.391 Right. You record signal from your muscle, from the patient. 00:15:30.631 --> 00:15:35.051 And let's imagine, for instance, the need for a prosthesis. Yes. 00:15:35.761 --> 00:15:40.401 Which should be equipped with actuators in order to be an active prosthesis. 00:15:40.581 --> 00:15:49.361 And you have to control the device, possibly using the physiological signal 00:15:49.361 --> 00:15:52.401 captured from the patient himself. 00:15:53.201 --> 00:15:59.541 So in this case, you capture muscular activity from any part of the body. 00:15:59.601 --> 00:16:03.041 For instance, if we are talking about a prosthetic hand, maybe from the forearm. 00:16:03.041 --> 00:16:06.821 Arm and you capture this this 00:16:06.821 --> 00:16:13.081 electrophysiological signal and with a very simple elaboration in real time 00:16:13.081 --> 00:16:20.001 you can drive your actuator in order to perform a specific task this is feasible 00:16:20.001 --> 00:16:25.381 also with other actuation technologies i don't see any difference with other 00:16:25.381 --> 00:16:26.681 technologies from this respect. 00:16:27.201 --> 00:16:32.681 In any case, we have confirmed that even using electrostatic polymer, 00:16:32.841 --> 00:16:35.501 this can be done, and this is important, of course. 00:16:35.961 --> 00:16:42.341 SL. But you think there are some particular areas where this technology might be much better? 00:16:42.701 --> 00:16:48.601 DR. Definitely. There are at least four emerging areas where this technology 00:16:48.601 --> 00:16:53.741 shows a great potential over any other attrition technologies available today. 00:16:53.921 --> 00:17:00.921 The first area is mechatronics, and in particular variable stiffness devices. 00:17:01.481 --> 00:17:06.981 For instance, in the biomedical field, we are talking about variable stiffness 00:17:06.981 --> 00:17:10.721 system for rehabilitation. 00:17:12.261 --> 00:17:18.401 We are developing a so-called hand splint orthosis in order to perform. 00:17:20.701 --> 00:17:26.061 Customized rehabilitation habilitation of the hand for post-stroke patients. 00:17:26.881 --> 00:17:32.881 So this is a big field of interest. So an orthosis is a support for the hand 00:17:32.881 --> 00:17:34.941 while it's healing? Right, it is a support. 00:17:35.841 --> 00:17:40.441 And in this case, hand orthosis are called splint, hand splint. 00:17:40.541 --> 00:17:44.121 Basically, they are used to, they are dynamic. 00:17:45.281 --> 00:17:49.901 That dynamic devices in the sense that they are equipped with elastic bands 00:17:49.901 --> 00:17:59.421 or springs in order to allow the patient to voluntarily move his fingers against definite loads, 00:17:59.661 --> 00:18:04.841 counter loads provided by the springs or the elastic bands in order to perform 00:18:04.841 --> 00:18:07.281 an exercise of their fingers. 00:18:08.332 --> 00:18:15.532 For instance, for this application, we are developing an active version of this 00:18:15.532 --> 00:18:22.072 passive system of the state of the art by replacing the springs or the elastic 00:18:22.072 --> 00:18:25.592 band with these elastic actuators. 00:18:25.592 --> 00:18:33.392 So the future splint will not be equipped anymore with passive springs or plastic 00:18:33.392 --> 00:18:35.652 band, but will have active bands. 00:18:35.772 --> 00:18:43.612 Active in the sense that by using an electrical control, you can modulate the stiffness of your band. 00:18:43.612 --> 00:18:50.072 So you can modulate the compliance, and so the patient can really perform a 00:18:50.072 --> 00:18:54.992 customized training against controllable loads. 00:18:55.172 --> 00:19:03.052 So the load is no more predefined a priori, but can be adapted in progress to 00:19:03.052 --> 00:19:05.132 the specific need of the patient. 00:19:06.212 --> 00:19:11.292 And this is important. So can you explain a bit how you can vary the stiffness? 00:19:11.292 --> 00:19:16.972 So I can see that when you take this polymer and you apply the current, 00:19:17.032 --> 00:19:19.892 you can change the shape, but how do you change the stiffness? 00:19:20.172 --> 00:19:24.552 It depends on how you basically, the specific constraint that you, 00:19:24.632 --> 00:19:27.372 how you use your material and your device in general. 00:19:27.492 --> 00:19:32.032 It is a little bit technical, this part, maybe it is difficult. 00:19:32.332 --> 00:19:33.332 But the question would be. 00:19:35.293 --> 00:19:40.053 Can you choose any level of force and at any level of stiffness, 00:19:40.153 --> 00:19:43.453 or is there some relationship between those two things? 00:19:43.453 --> 00:19:48.253 We have demonstrated, but this has been anticipated also by theoretical calculation, 00:19:48.633 --> 00:19:53.433 we have demonstrated experimentally that you can really control the stiffness. 00:19:54.293 --> 00:20:01.693 So no more a simple position or force control, but a real stiffness control. 00:20:01.693 --> 00:20:06.253 You can really control the stiffness, which is basically the slope of the force 00:20:06.253 --> 00:20:07.893 versus the displacement. 00:20:08.313 --> 00:20:12.393 At least what we call the static stiffness. 00:20:13.813 --> 00:20:17.133 And this can be done. We have demonstrated. So it is important for, 00:20:17.233 --> 00:20:21.373 as I said, for instance, for biomedical application oriented to rehabilitation. 00:20:21.953 --> 00:20:26.253 So if you can control the stiffness, then there are lots of applications in 00:20:26.253 --> 00:20:30.653 virtual reality where you might be able to use this as a display. You're right. 00:20:31.353 --> 00:20:39.293 But maybe not always. For instance, imagine about vibration damping in vehicles. 00:20:39.873 --> 00:20:44.933 There are many activities going on, especially in Germany, where there are a 00:20:44.933 --> 00:20:47.173 lot of car manufacturers interested 00:20:47.173 --> 00:20:52.713 in using this technology for vibration damping in cars, for instance. 00:20:52.713 --> 00:21:03.273 So if it can control the stiffness, let's say of some parts of the interior of the car, 00:21:04.033 --> 00:21:09.533 someone is even thinking about something more challenging, maybe the motor itself, 00:21:09.893 --> 00:21:11.873 then you can really damp vibration. 00:21:11.953 --> 00:21:16.453 This can be useful for the vehicle, but also for the passengers in order to 00:21:16.453 --> 00:21:17.973 make the travel smoother. 00:21:18.513 --> 00:21:21.413 By the way, this is a work in progress. Yeah. 00:21:21.553 --> 00:21:26.473 Can you give me an example of where you might use this material to display something 00:21:26.473 --> 00:21:31.593 that I might feel on the skin and to represent that? 00:21:32.386 --> 00:21:35.926 Or to change the property of that surface that I experience through touch, 00:21:36.046 --> 00:21:39.186 through haptic. Yeah, so you're talking about haptic devices. Yes. 00:21:39.686 --> 00:21:45.566 That's another interesting area where there is significant potential. 00:21:45.626 --> 00:21:49.846 For instance, we are developing a braille display. 00:21:50.386 --> 00:21:53.706 So a tactile display for the blind people. 00:21:54.886 --> 00:21:58.126 The state of the art of this display is the following. 00:21:58.126 --> 00:22:06.226 Basically, you have some dots, plastic dots, which go up and down and form the 00:22:06.226 --> 00:22:09.986 Braille character, so the Braille code for the blind people. 00:22:10.946 --> 00:22:15.626 And the unit Braille cell consists of eight dots. 00:22:15.846 --> 00:22:21.346 By changing the status of these eight dots, different combinations provide different characters. 00:22:21.346 --> 00:22:24.746 So in these devices the 00:22:24.746 --> 00:22:28.866 state-of-the-art devices are driven by cantilever 00:22:28.866 --> 00:22:31.686 like piezoelectric actuators which are 00:22:31.686 --> 00:22:35.526 very long let's say a long rigid shaft which 00:22:35.526 --> 00:22:38.566 pull push these dots up 00:22:38.566 --> 00:22:44.306 and down and the problem is that the the encumbrance of these devices is very 00:22:44.306 --> 00:22:52.886 they are very bulky So you cannot today implement more than two reading rows 00:22:52.886 --> 00:22:55.926 because of the volume constraint. 00:22:56.346 --> 00:23:03.666 That's it. So Braille displays as a full page representation of the Braille 00:23:03.666 --> 00:23:07.526 code are not achievable today with the state-of-the-art technology. 00:23:07.526 --> 00:23:13.186 So, in order to overcome this limitation, let's say try to develop a sort of 00:23:13.186 --> 00:23:18.226 electronic book for the blind people, as we have in the tablet nowadays, 00:23:18.746 --> 00:23:23.326 you really need something, a technology really more compact. 00:23:24.086 --> 00:23:28.906 And for instance, we are developing now with the help of a company, 00:23:29.026 --> 00:23:35.206 we are developing some small dots, braille dots, and each dot itself is both 00:23:35.206 --> 00:23:37.206 the actuator and the braille dot. 00:23:37.426 --> 00:23:43.466 So we have tiny bubbles that go up and down, and we hope that we will be able 00:23:43.466 --> 00:23:47.086 to demonstrate a new braille display soon. 00:23:48.104 --> 00:23:53.544 And you could imagine, as a generalization of that, that you could take surfaces 00:23:53.544 --> 00:23:57.404 and you could change in real time their haptic properties, 00:23:57.664 --> 00:24:01.324 so you could take something that's smooth and it could become ribbed, 00:24:01.364 --> 00:24:04.304 and there could be all sorts of interesting applications for that. 00:24:04.404 --> 00:24:09.504 The fathers of this technology in the US from the Stanford Research Institute, 00:24:10.344 --> 00:24:17.084 some years ago, have already demonstrated some prototype with texture changing surfaces. 00:24:18.104 --> 00:24:21.824 Using this concept, well, not exactly this concept, but something close, 00:24:21.944 --> 00:24:23.504 but by the way, using this technology. 00:24:23.824 --> 00:24:28.484 So in the future, you might imagine something that was like a sheet, 00:24:28.624 --> 00:24:33.364 which would deform itself to form a tactile picture of an object. 00:24:33.584 --> 00:24:38.664 That is a dream, but it is realistically achievable from our point of view. 00:24:38.784 --> 00:24:43.524 It is a matter of trying to miniaturize and, you know, the technology improve 00:24:43.524 --> 00:24:47.284 the resolution, but it is technically feasible. 00:24:48.104 --> 00:24:54.344 I would like to add a comment about other possible areas. We said mechatronics. 00:24:54.804 --> 00:24:59.584 Another big area is energy harvesting. So this would be where you drive it in reverse? 00:24:59.964 --> 00:25:01.304 Exactly, in reverse. You can 00:25:01.304 --> 00:25:06.764 generate electricity by harvesting mechanical energy from the exterior. 00:25:07.024 --> 00:25:11.144 So if you deform this material when they are electrically charged, 00:25:11.284 --> 00:25:15.864 and then you release the force that you have applied to the material to deform 00:25:15.864 --> 00:25:19.564 them, then the electrical charge stored in the material. 00:25:21.204 --> 00:25:24.144 Increases the electrical energy, 00:25:25.122 --> 00:25:28.922 This can be easily demonstrated with calculation, but also with experimental tests. 00:25:29.202 --> 00:25:35.002 So basically, you convert the input mechanical energy into output electrical energy. 00:25:35.342 --> 00:25:39.022 And for instance, there are a lot of studies nowadays and also experimental 00:25:39.022 --> 00:25:45.162 tests to harvest the mechanical energy from the ocean waves. 00:25:45.722 --> 00:25:50.662 So this is not too different from an existing electric dynamo, presumably. 00:25:51.062 --> 00:25:56.442 Exactly. Somehow the concept is the same. but the principle of operation is different. 00:25:56.762 --> 00:26:05.242 And this is very useful nowadays because we all know that worldwide energy consumption is increasing. 00:26:05.322 --> 00:26:09.122 We need renewable energy sources. 00:26:09.482 --> 00:26:16.062 And for instance, if we imagine the sea waves are an enormous provider of free 00:26:16.062 --> 00:26:20.622 energy, it is just a matter of finding really the best technologies. 00:26:21.442 --> 00:26:24.682 Able to harvest this mechanical energy. 00:26:25.182 --> 00:26:29.482 Similarly, we can harvest energy from the wind using the same technology. 00:26:29.882 --> 00:26:33.782 Is the idea that this would be more energy efficient and that's largely a consequence 00:26:33.782 --> 00:26:37.402 of it? I would say for sure that will be much more cost effective because you 00:26:37.402 --> 00:26:40.942 can produce very large surfaces at low cost. 00:26:41.062 --> 00:26:44.062 And this is not feasible with any other technology because 00:26:44.062 --> 00:26:46.982 this technology is polymer based and 00:26:46.982 --> 00:26:50.322 polymer are very cost effective 00:26:50.322 --> 00:26:56.762 material so this is a great advantage moreover they work nicely at low frequencies 00:26:56.762 --> 00:27:05.202 especially at low frequencies so they are nicely complementary to conventional 00:27:05.202 --> 00:27:09.442 technologies based for instance on piezoelectric energy harvester, 00:27:09.582 --> 00:27:13.422 which nicely work at resonance at higher frequencies. 00:27:14.598 --> 00:27:19.718 So this is another interesting field. So another thing that you were mentioning 00:27:19.718 --> 00:27:23.398 is that you're taking biological inspiration from another place, 00:27:23.458 --> 00:27:27.078 which is from how humans use their eyes. Exactly. 00:27:27.698 --> 00:27:33.138 Tunable optics is the last really emerging field of application. 00:27:33.798 --> 00:27:39.978 And we are quite happy because we have recently demonstrated the first tunable 00:27:39.978 --> 00:27:46.538 lens aspire to the architecture of the crystalline lens in our human eye. 00:27:46.758 --> 00:27:53.798 So basically in our eye we have a lens which is deformed by the action of some 00:27:53.798 --> 00:28:00.878 muscles which are called ciliary muscle which basically stretch and release 00:28:00.878 --> 00:28:06.678 the crystalline lens in order to shape its this curvature, 00:28:07.018 --> 00:28:10.518 and then in order to change the focus. 00:28:10.758 --> 00:28:15.178 So the lens is a bit like a soft ball that you can pull it one direction, 00:28:15.258 --> 00:28:19.078 make it flatter, and squash it and make it more spherical. 00:28:19.138 --> 00:28:23.558 Exactly. And we have some dedicated muscle for this action in our human eye. 00:28:24.658 --> 00:28:29.658 Similarly, we have produced two membranes filled with a fluid, 00:28:30.278 --> 00:28:34.178 surrounded by one of these dielectric elastomer actuators, 00:28:34.478 --> 00:28:37.758 everything has been developed in our lab and 00:28:37.758 --> 00:28:43.818 the actuator as a real artificial muscle is able to deform the lens radially 00:28:43.818 --> 00:28:50.158 so to change the focus and we have demonstrated as you have seen at this school 00:28:50.158 --> 00:28:57.158 also live i have brought a demo you have seen how nicely it is able to to really change the focus. 00:28:57.378 --> 00:29:06.058 So we expect that this biomimetic approach will really provide some new. 00:29:06.938 --> 00:29:10.058 Capabilities to tunable lenses in order 00:29:10.058 --> 00:29:13.898 to make them really compact lightweight again 00:29:13.898 --> 00:29:17.278 power efficient and very 00:29:17.278 --> 00:29:21.838 cheap at the same time and could you imagine a future prosthetic eye which was 00:29:21.838 --> 00:29:26.398 using this that's what that's our dream actually we are working for for a prosthetic 00:29:26.398 --> 00:29:32.098 eye based on this technology so all these prosthetics are amazing but one thing 00:29:32.098 --> 00:29:34.018 that i've heard about these EAPs, 00:29:34.038 --> 00:29:37.938 is that you have to put an enormous voltage across the surface. 00:29:38.138 --> 00:29:43.238 And that sounds quite terrifying. It is terrifying, but if you don't know the 00:29:43.238 --> 00:29:45.758 full story, let me make a remark on that. 00:29:45.938 --> 00:29:51.378 It is true that you need high voltages, today in the order of one kilovolt, 00:29:51.378 --> 00:29:58.978 because we have materials that are completely not optimized for this purpose. 00:29:58.978 --> 00:30:03.098 Nowadays, we use some, for instance, silicones, which have been developed by 00:30:03.098 --> 00:30:04.558 industry for other purposes. 00:30:05.478 --> 00:30:13.698 So these materials do not have specific electrical properties tailored for this application. 00:30:13.898 --> 00:30:16.518 In particular, they do not have high dielectric constant. 00:30:16.858 --> 00:30:22.698 And the higher is the dielectric constant, the lower is the electric field that you have to apply. 00:30:22.918 --> 00:30:26.538 And so the lower is the voltage that you have to apply. So this is a problem 00:30:26.538 --> 00:30:28.238 related to material science. 00:30:28.978 --> 00:30:33.998 That is now faced by several groups because it is really evident that if we 00:30:33.998 --> 00:30:37.698 want to reduce the driving voltage we have to increase that electric constant. 00:30:38.689 --> 00:30:45.409 On the technological side, however, we can still play with available off-the-shelf 00:30:45.409 --> 00:30:50.029 material by simply manufacturing thin layers. 00:30:50.349 --> 00:30:55.349 The thinner is the layer, the lower is the voltage that you have to apply. 00:30:55.569 --> 00:31:00.509 Because the important parameter, important variable, is the electric field, 00:31:00.649 --> 00:31:05.189 which is the ratio between the applied voltage and the thickness of the layer. 00:31:05.329 --> 00:31:08.429 So lower thickness, lower voltages. 00:31:08.689 --> 00:31:11.549 And in principle but also in 00:31:11.549 --> 00:31:18.129 practice you can drive these materials with voltages of the order of 100 volts 00:31:18.129 --> 00:31:23.809 this has been already demonstrated by many groups so this is not a major issue 00:31:23.809 --> 00:31:29.169 for the future but will it require special ways of manufacturing to make these things? 00:31:29.729 --> 00:31:37.809 Yes, using voltages of the order of 100 volts which is the same order of Manitou-Piezio 00:31:37.809 --> 00:31:40.929 electrics It's not so dramatically difficult. 00:31:41.309 --> 00:31:46.809 It can be done even by university lab, not only by companies. 00:31:46.949 --> 00:31:50.909 The problem is that if you want to drive them at 10 volt, 00:31:51.209 --> 00:31:58.529 that is really challenging because you need an order of one micron thick film, 00:31:58.669 --> 00:32:02.489 which is challenging with this material because they are extremely soft. 00:32:02.649 --> 00:32:07.709 So in the long future, certainly this is the direction to be faced, 00:32:07.709 --> 00:32:13.849 but we still have to see, I mean, the real possibilities from that point of view. 00:32:13.909 --> 00:32:18.009 But I have to remark that it is true that this can be a challenge, 00:32:18.149 --> 00:32:24.369 but not a critical point from a technical standpoint, because you can easily 00:32:24.369 --> 00:32:29.129 produce these high voltages with very compact multipliers. 00:32:30.497 --> 00:32:35.177 I mean, very compact, I say a few millimetre cube, cube millimetres. 00:32:35.217 --> 00:32:39.397 So very, very simple electronic components can generate these high voltages 00:32:39.397 --> 00:32:43.037 because you don't need high current. And this is the key point. 00:32:43.217 --> 00:32:47.877 You need to drive this technology with high voltages, but you don't need to 00:32:47.877 --> 00:32:52.637 supply high current because these loads are capacitive loads. 00:32:52.917 --> 00:32:57.537 So they do not absorb high current like a resistive load. 00:32:57.737 --> 00:33:00.817 And this is a winning point. So even if you get a shock off this, 00:33:00.897 --> 00:33:04.537 it wouldn't kill you? Yeah, even if you get a shock, it is unpleasant, 00:33:04.857 --> 00:33:09.097 like a shock that you get from your car, but it is not dangerous. 00:33:10.357 --> 00:33:14.057 That's very important. And that's why, for instance, I mentioned before, 00:33:14.237 --> 00:33:16.897 a big company has been able to deliver 00:33:16.897 --> 00:33:21.817 on the market the first portable application for a mobile phone. on. 00:33:21.997 --> 00:33:26.017 This has been possible because the current that you have to supply is small. 00:33:26.177 --> 00:33:32.677 So you can simply use a battery like the battery used in our cell phone today. 00:33:33.137 --> 00:33:39.317 It is not dangerous. So I mean, it is allowed for commercial product to play 00:33:39.317 --> 00:33:41.617 at one kilovolt with no risk. 00:33:41.777 --> 00:33:46.057 So it is true that maybe for the men of the street, this can be a scaring factor, 00:33:46.217 --> 00:33:49.357 but actually it is not a technical problem nowadays 00:33:49.357 --> 00:33:52.777 so we'll be carrying around these things in our pockets oh 00:33:52.777 --> 00:33:55.577 yeah one kilovolt yes and they won't be able to do oh yes if you 00:33:55.577 --> 00:33:58.497 if you go to the market tomorrow you still can buy 00:33:58.497 --> 00:34:02.997 you already can buy okay one of those mobile phones so you will have one kilovolt 00:34:02.997 --> 00:34:08.437 in your pocket so in 10 years time yeah what do you predict uh what's one of 00:34:08.437 --> 00:34:13.217 your predictions as to a device that i will be using perhaps every day oh i 00:34:13.217 --> 00:34:15.817 predict there will be really an explosion of this technology, 00:34:16.057 --> 00:34:19.197 because I've seen that from the last 10 years. 00:34:19.357 --> 00:34:27.157 10 years ago, the first paper on science, on the fundamentally of this technology 00:34:27.157 --> 00:34:29.237 was published, really 10 years ago. 00:34:29.357 --> 00:34:35.957 In 10 years, we have achieved the first industrial product in a mobile phone. 00:34:36.077 --> 00:34:40.257 And 10 years for a new technology are really nothing. 00:34:40.397 --> 00:34:48.177 So it means that the technology is really promising. is really suitable for great developments. 00:34:48.497 --> 00:34:50.577 And I expect that the next 10 00:34:50.577 --> 00:34:55.917 years will provide a great evidence of the importance of this technology. 00:34:57.882 --> 00:35:01.542 There was a specific technology that you could mention that you'd be really 00:35:01.542 --> 00:35:05.402 excited about having in 10 years' time. What would that be that would use this? 00:35:05.802 --> 00:35:11.202 I mean the specific product? Yeah, yeah. Well, for sure, optical devices and 00:35:11.202 --> 00:35:16.262 haptic devices will be the major fields of application. 00:35:16.602 --> 00:35:24.242 So, for instance, tunable lenses and tactile, vibrotactile devices will have a great, great role. 00:35:24.562 --> 00:35:27.802 So, when you say vibrotactile devices, you're 00:35:27.802 --> 00:35:31.042 thinking of something like the nintendo wii controller 00:35:31.042 --> 00:35:34.162 which now buzzes in your hand when 00:35:34.162 --> 00:35:38.342 you're playing a game for instance this is the current application in 10 years 00:35:38.342 --> 00:35:43.302 time what will that be doing that that vibrotactile device well as a biomedical 00:35:43.302 --> 00:35:48.802 engineer i have a dream i hope that in 10 years at least some of our force will 00:35:48.802 --> 00:35:51.062 be will have been useful for the blind people. 00:35:51.202 --> 00:35:58.222 This is my greatest wish because we will really enable something which is not possible today. 00:35:58.482 --> 00:36:01.942 So a great advantage for the blind people. 00:36:02.282 --> 00:36:05.602 Okay. Well, thanks very much for talking to us. It's really interesting. Thanks. 00:36:09.722 --> 00:36:15.302 The CSN podcast was produced by the Convergent Science Network of Biometrics 00:36:15.302 --> 00:36:22.122 and Biohybrid A project funded by the European Sevens Research Framework Programme. 00:36:23.122 --> 00:36:28.582 For more interviews, recorded lectures or upcoming conferences in the field 00:36:28.582 --> 00:36:34.662 of biometrics and biohybrid systems, go to csnnetwork.com. 00:36:35.120 --> 00:36:42.960 Music.