WEBVTT 00:00:00.760 --> 00:00:02.600 This is Pleurobot. 00:00:03.400 --> 00:00:07.016 Pleurobot is a robot that we designed to closely mimic a salamander species 00:00:07.040 --> 00:00:08.440 called Pleurodeles waltl. 00:00:09.240 --> 00:00:11.496 Pleurobot can walk, as you can see here, 00:00:11.520 --> 00:00:13.560 and as you'll see later, it can also swim. NOTE Paragraph 00:00:14.280 --> 00:00:16.471 So you might ask, why did we design this robot? 00:00:16.960 --> 00:00:20.722 And in fact, this robot has been designed as a scientific tool for neuroscience. 00:00:21.400 --> 00:00:23.896 Indeed, we designed it together with neurobiologists 00:00:23.920 --> 00:00:25.816 to understand how animals move, 00:00:25.840 --> 00:00:28.600 and especially how the spinal cord controls locomotion. 00:00:29.560 --> 00:00:31.256 But the more I work in biorobotics, 00:00:31.280 --> 00:00:33.661 the more I'm really impressed by animal locomotion. 00:00:33.920 --> 00:00:38.216 If you think of a dolphin swimming or a cat running or jumping around, 00:00:38.240 --> 00:00:39.816 or even us as humans, 00:00:39.840 --> 00:00:41.656 when you go jogging or play tennis, 00:00:41.680 --> 00:00:42.920 we do amazing things. 00:00:43.880 --> 00:00:48.016 And in fact, our nervous system solves a very, very complex control problem. 00:00:48.040 --> 00:00:51.136 It has to coordinate more or less 200 muscles perfectly, 00:00:51.160 --> 00:00:54.840 because if the coordination is bad, we fall over or we do bad locomotion. 00:00:55.560 --> 00:00:58.280 And my goal is to understand how this works. NOTE Paragraph 00:00:59.160 --> 00:01:02.000 There are four main components behind animal locomotion. 00:01:02.800 --> 00:01:04.736 The first component is just the body, 00:01:04.760 --> 00:01:06.736 and in fact we should never underestimate 00:01:06.760 --> 00:01:10.240 to what extent the biomechanics already simplify locomotion in animals. 00:01:10.920 --> 00:01:12.376 Then you have the spinal cord, 00:01:12.400 --> 00:01:14.376 and in the spinal cord you find reflexes, 00:01:14.400 --> 00:01:17.856 multiple reflexes that create a sensorimotor coordination loop 00:01:17.880 --> 00:01:21.360 between neural activity in the spinal cord and mechanical activity. 00:01:22.000 --> 00:01:24.976 A third component are central pattern generators. 00:01:25.000 --> 00:01:28.896 These are very interesting circuits in the spinal cord of vertebrate animals 00:01:28.920 --> 00:01:30.536 that can generate, by themselves, 00:01:30.560 --> 00:01:33.296 very coordinated rhythmic patterns of activity 00:01:33.320 --> 00:01:35.696 while receiving only very simple input signals. 00:01:35.720 --> 00:01:36.936 And these input signals 00:01:36.960 --> 00:01:40.016 coming from descending modulation from higher parts of the brain, 00:01:40.040 --> 00:01:42.736 like the motor cortex, the cerebellum, the basal ganglia, 00:01:42.760 --> 00:01:44.896 will all modulate activity of the spinal cord 00:01:44.920 --> 00:01:46.376 while we do locomotion. 00:01:46.400 --> 00:01:49.616 But what's interesting is to what extent just a low-level component, 00:01:49.640 --> 00:01:51.576 the spinal cord, together with the body, 00:01:51.600 --> 00:01:54.056 already solve a big part of the locomotion problem. 00:01:54.080 --> 00:01:57.502 You probably know it by the fact that you can cut the head off a chicken, 00:01:57.532 --> 00:01:58.913 it can still run for a while, 00:01:58.937 --> 00:02:01.476 showing that just the lower part, spinal cord and body, 00:02:01.500 --> 00:02:03.373 already solve a big part of locomotion. NOTE Paragraph 00:02:03.397 --> 00:02:05.856 Now, understanding how this works is very complex, 00:02:05.880 --> 00:02:07.176 because first of all, 00:02:07.200 --> 00:02:09.820 recording activity in the spinal cord is very difficult. 00:02:09.844 --> 00:02:12.616 It's much easier to implant electrodes in the motor cortex 00:02:12.640 --> 00:02:15.696 than in the spinal cord, because it's protected by the vertebrae. 00:02:15.720 --> 00:02:17.536 Especially in humans, very hard to do. 00:02:17.560 --> 00:02:21.336 A second difficulty is that locomotion is really due to a very complex 00:02:21.360 --> 00:02:24.416 and very dynamic interaction between these four components. 00:02:24.440 --> 00:02:27.680 So it's very hard to find out what's the role of each over time. 00:02:28.880 --> 00:02:32.616 This is where biorobots like Pleurobot and mathematical models 00:02:32.640 --> 00:02:33.840 can really help. NOTE Paragraph 00:02:35.480 --> 00:02:36.736 So what's biorobotics? 00:02:36.760 --> 00:02:39.496 Biorobotics is a very active field of research in robotics 00:02:39.520 --> 00:02:41.976 where people want to take inspiration from animals 00:02:42.000 --> 00:02:44.456 to make robots to go outdoors, 00:02:44.480 --> 00:02:47.136 like service robots or search and rescue robots 00:02:47.160 --> 00:02:48.360 or field robots. 00:02:48.880 --> 00:02:51.576 And the big goal here is to take inspiration from animals 00:02:51.600 --> 00:02:53.936 to make robots that can handle complex terrain -- 00:02:53.960 --> 00:02:55.576 stairs, mountains, forests, 00:02:55.600 --> 00:02:57.616 places where robots still have difficulties 00:02:57.640 --> 00:02:59.696 and where animals can do a much better job. 00:02:59.720 --> 00:03:02.216 The robot can be a wonderful scientific tool as well. 00:03:02.240 --> 00:03:04.860 There are some very nice projects where robots are used, 00:03:04.884 --> 00:03:08.856 like a scientific tool for neuroscience, for biomechanics or for hydrodynamics. 00:03:08.880 --> 00:03:11.000 And this is exactly the purpose of Pleurobot. 00:03:11.600 --> 00:03:14.536 So what we do in my lab is to collaborate with neurobiologists 00:03:14.560 --> 00:03:17.776 like Jean-Marie Cabelguen, a neurobiologist in Bordeaux in France, 00:03:17.800 --> 00:03:21.840 and we want to make spinal cord models and validate them on robots. 00:03:22.480 --> 00:03:24.096 And here we want to start simple. NOTE Paragraph 00:03:24.120 --> 00:03:26.096 So it's good to start with simple animals 00:03:26.120 --> 00:03:28.376 like lampreys, which are very primitive fish, 00:03:28.400 --> 00:03:30.896 and then gradually go toward more complex locomotion, 00:03:30.920 --> 00:03:32.176 like in salamanders, 00:03:32.200 --> 00:03:33.696 but also in cats and in humans, 00:03:33.720 --> 00:03:34.920 in mammals. 00:03:35.880 --> 00:03:38.256 And here, a robot becomes an interesting tool 00:03:38.280 --> 00:03:40.216 to validate our models. 00:03:40.240 --> 00:03:43.256 And in fact, for me, Pleurobot is a kind of dream becoming true. 00:03:43.280 --> 00:03:46.536 Like, more or less 20 years ago I was already working on a computer 00:03:46.560 --> 00:03:49.216 making simulations of lamprey and salamander locomotion 00:03:49.240 --> 00:03:50.776 during my PhD. 00:03:50.800 --> 00:03:54.176 But I always knew that my simulations were just approximations. 00:03:54.200 --> 00:03:58.176 Like, simulating the physics in water or with mud or with complex ground, 00:03:58.200 --> 00:04:00.856 it's very hard to simulate that properly on a computer. 00:04:00.880 --> 00:04:02.920 Why not have a real robot and real physics? 00:04:03.600 --> 00:04:06.736 So among all these animals, one of my favorites is the salamander. 00:04:06.760 --> 00:04:10.216 You might as why, and it's because as an amphibian, 00:04:10.240 --> 00:04:13.096 it's a really key animal from an evolutionary point of view. 00:04:13.120 --> 00:04:15.176 It makes a wonderful link between swimming, 00:04:15.200 --> 00:04:17.096 as you find it in eels or fish, 00:04:17.120 --> 00:04:21.240 and quadruped locomotion, as you see in mammals, in cats and humans. 00:04:22.160 --> 00:04:23.816 And in fact, the modern salamander 00:04:23.840 --> 00:04:26.216 is very close to the first terrestrial vertebrate, 00:04:26.240 --> 00:04:27.776 so it's almost a living fossil, 00:04:27.800 --> 00:04:29.736 which gives us access to our ancestor, 00:04:29.760 --> 00:04:32.680 the ancestor to all terrestrial tetrapods. NOTE Paragraph 00:04:33.240 --> 00:04:34.616 So the salamander swims 00:04:34.640 --> 00:04:37.136 by doing what's called an anguilliform swimming gait, 00:04:37.160 --> 00:04:40.800 so they propagate a nice traveling wave of muscle activity from head to tail. 00:04:41.440 --> 00:04:43.616 And if you place the salamander on the ground, 00:04:43.640 --> 00:04:45.976 it switches to what's called a walking trot gait. 00:04:46.000 --> 00:04:48.863 In this case, you have nice periodic activation of the limbs 00:04:48.887 --> 00:04:50.496 which are very nicely coordinated 00:04:50.520 --> 00:04:53.176 with this standing wave undulation of the body, 00:04:53.200 --> 00:04:56.856 and that's exactly the gait that you are seeing here on Pleurobot. 00:04:56.880 --> 00:04:59.856 Now, one thing which is very surprising and fascinating in fact 00:04:59.880 --> 00:05:04.016 is the fact that all this can be generated just by the spinal cord and the body. 00:05:04.040 --> 00:05:06.040 So if you take a decerebrated salamander -- 00:05:06.064 --> 00:05:08.080 it's not so nice but you remove the head -- 00:05:08.104 --> 00:05:10.776 and if you electrically stimulate the spinal cord, 00:05:10.800 --> 00:05:14.056 at low level of stimulation this will induce a walking-like gait. 00:05:14.080 --> 00:05:16.536 If you stimulate a bit more, the gait accelerates. 00:05:16.560 --> 00:05:18.456 And at some point, there's a threshold, 00:05:18.480 --> 00:05:21.016 and automatically, the animal switches to swimming. 00:05:21.040 --> 00:05:22.416 This is amazing. 00:05:22.440 --> 00:05:23.936 Just changing the global drive, 00:05:23.960 --> 00:05:25.696 as if you are pressing the gas pedal 00:05:25.720 --> 00:05:27.856 of descending modulation to your spinal cord, 00:05:27.880 --> 00:05:30.880 makes a complete switch between two very different gaits. 00:05:32.440 --> 00:05:35.016 And in fact, the same has been observed in cats. 00:05:35.040 --> 00:05:37.056 If you stimulate the spinal cord of a cat, 00:05:37.080 --> 00:05:39.296 you can switch between walk, trot and gallop. 00:05:39.320 --> 00:05:42.056 Or in birds, you can make a bird switch between walking, 00:05:42.080 --> 00:05:43.536 at a low level of stimulation, 00:05:43.560 --> 00:05:46.376 and flapping its wings at high-level stimulation. 00:05:46.400 --> 00:05:48.416 And this really shows that the spinal cord 00:05:48.440 --> 00:05:50.856 is a very sophisticated locomotion controller. NOTE Paragraph 00:05:50.880 --> 00:05:53.336 So we studied salamander locomotion in more detail, 00:05:53.360 --> 00:05:56.456 and we had in fact access to a very nice X-ray video machine 00:05:56.480 --> 00:06:00.056 from Professor Martin Fischer in Jena University in Germany. 00:06:00.080 --> 00:06:02.656 And thanks to that, you really have an amazing machine 00:06:02.680 --> 00:06:05.136 to record all the bone motion in great detail. 00:06:05.160 --> 00:06:06.416 That's what we did. 00:06:06.440 --> 00:06:09.616 So we basically figured out which bones are important for us 00:06:09.640 --> 00:06:12.656 and collected their motion in 3D. 00:06:12.680 --> 00:06:15.376 And what we did is collect a whole database of motions, 00:06:15.400 --> 00:06:17.056 both on ground and in water, 00:06:17.080 --> 00:06:19.564 to really collect a whole database of motor behaviors 00:06:19.589 --> 00:06:20.833 that a real animal can do. 00:06:20.858 --> 00:06:24.008 And then our job as roboticists was to replicate that in our robot. 00:06:24.033 --> 00:06:27.416 So we did a whole optimization process to find out the right structure, 00:06:27.440 --> 00:06:30.096 where to place the motors, how to connect them together, 00:06:30.120 --> 00:06:33.000 to be able to replay these motions as well as possible. 00:06:33.680 --> 00:06:36.040 And this is how Pleurobot came to life. NOTE Paragraph 00:06:37.200 --> 00:06:39.616 So let's look at how close it is to the real animal. 00:06:40.960 --> 00:06:43.456 So what you see here is almost a direct comparison 00:06:43.480 --> 00:06:46.176 between the walking of the real animal and the Pleurobot. 00:06:46.200 --> 00:06:48.936 You can see that we have almost a one-to-one exact replay 00:06:48.960 --> 00:06:50.216 of the walking gait. 00:06:50.240 --> 00:06:52.840 If you go backwards and slowly, you see it even better. 00:06:55.520 --> 00:06:57.896 But even better, we can do swimming. 00:06:57.920 --> 00:07:00.936 So for that we have a dry suit that we put all over the robot -- NOTE Paragraph 00:07:00.960 --> 00:07:02.056 (Laughter) NOTE Paragraph 00:07:02.080 --> 00:07:05.256 and then we can go in water and start replaying the swimming gaits. 00:07:05.280 --> 00:07:08.616 And here, we were very happy, because this is difficult to do. 00:07:08.640 --> 00:07:10.856 The physics of interaction are complex. 00:07:10.880 --> 00:07:13.296 Our robot is much bigger than a small animal, 00:07:13.320 --> 00:07:16.376 so we had to do what's called dynamic scaling of the frequencies 00:07:16.400 --> 00:07:18.736 to make sure we had the same interaction physics. 00:07:18.760 --> 00:07:21.176 But you see at the end, we have a very close match, 00:07:21.200 --> 00:07:23.080 and we were very, very happy with this. 00:07:23.480 --> 00:07:25.696 So let's go to the spinal cord. 00:07:25.720 --> 00:07:28.016 So here what we did with Jean-Marie Cabelguen 00:07:28.040 --> 00:07:30.280 is model the spinal cord circuits. 00:07:31.040 --> 00:07:33.176 And what's interesting is that the salamander 00:07:33.200 --> 00:07:34.820 has kept a very primitive circuit, 00:07:34.844 --> 00:07:37.496 which is very similar to the one we find in the lamprey, 00:07:37.520 --> 00:07:39.496 this primitive eel-like fish, 00:07:39.520 --> 00:07:41.256 and it looks like during evolution, 00:07:41.280 --> 00:07:44.216 new neural oscillators have been added to control the limbs, 00:07:44.240 --> 00:07:45.656 to do the leg locomotion. 00:07:45.680 --> 00:07:47.856 And we know where these neural oscillators are 00:07:47.880 --> 00:07:50.136 but what we did was to make a mathematical model 00:07:50.160 --> 00:07:51.776 to see how they should be coupled 00:07:51.800 --> 00:07:54.736 to allow this transition between the two very different gaits. 00:07:54.760 --> 00:07:57.320 And we tested that on board of a robot. NOTE Paragraph 00:07:57.680 --> 00:07:58.880 And this is how it looks. 00:08:06.920 --> 00:08:09.936 So what you see here is a previous version of Pleurobot 00:08:09.960 --> 00:08:13.056 that's completely controlled by our spinal cord model 00:08:13.080 --> 00:08:14.680 programmed on board of the robot. 00:08:15.280 --> 00:08:16.496 And the only thing we do 00:08:16.520 --> 00:08:18.696 is send to the robot through a remote control 00:08:18.720 --> 00:08:21.216 the two descending signals it normally should receive 00:08:21.240 --> 00:08:22.840 from the upper part of the brain. 00:08:23.480 --> 00:08:26.176 And what's interesting is, by playing with these signals, 00:08:26.200 --> 00:08:29.000 we can completely control speed, heading and type of gait. 00:08:29.600 --> 00:08:30.816 For instance, 00:08:30.840 --> 00:08:34.416 when we stimulate at a low level, we have the walking gait, 00:08:34.440 --> 00:08:36.416 and at some point, if we stimulate a lot, 00:08:36.440 --> 00:08:38.600 very rapidly it switches to the swimming gait. 00:08:39.480 --> 00:08:41.696 And finally, we can also do turning very nicely 00:08:41.720 --> 00:08:45.240 by just stimulating more one side of the spinal cord than the other. 00:08:46.200 --> 00:08:47.816 And I think it's really beautiful 00:08:47.840 --> 00:08:50.096 how nature has distributed control 00:08:50.120 --> 00:08:52.976 to really give a lot of responsibility to the spinal cord 00:08:53.000 --> 00:08:56.656 so that the upper part of the brain doesn't need to worry about every muscle. 00:08:56.680 --> 00:08:59.216 It just has to worry about this high-level modulation, 00:08:59.240 --> 00:09:02.816 and it's really the job of the spinal cord to coordinate all the muscles. NOTE Paragraph 00:09:02.840 --> 00:09:06.360 So now let's go to cat locomotion and the importance of biomechanics. 00:09:07.080 --> 00:09:08.336 So this is another project 00:09:08.360 --> 00:09:10.776 where we studied cat biomechanics, 00:09:10.800 --> 00:09:14.696 and we wanted to see how much the morphology helps locomotion. 00:09:14.720 --> 00:09:18.336 And we found three important criteria in the properties, 00:09:18.360 --> 00:09:19.680 basically, of the limbs. 00:09:20.320 --> 00:09:22.296 The first one is that a cat limb 00:09:22.320 --> 00:09:25.016 more or less looks like a pantograph-like structure. 00:09:25.040 --> 00:09:27.256 So a pantograph is a mechanical structure 00:09:27.280 --> 00:09:30.680 which keeps the upper segment and the lower segments always parallel. 00:09:31.600 --> 00:09:34.696 So a simple geometrical system that kind of coordinates a bit 00:09:34.720 --> 00:09:36.536 the internal movement of the segments. 00:09:36.560 --> 00:09:39.616 A second property of cat limbs is that they are very lightweight. 00:09:39.640 --> 00:09:41.496 Most of the muscles are in the trunk, 00:09:41.520 --> 00:09:44.416 which is a good idea, because then the limbs have low inertia 00:09:44.440 --> 00:09:46.216 and can be moved very rapidly. 00:09:46.240 --> 00:09:50.056 The last final important property is this very elastic behavior of the cat limb, 00:09:50.080 --> 00:09:52.736 so to handle impacts and forces. 00:09:52.760 --> 00:09:55.096 And this is how we designed Cheetah-Cub. NOTE Paragraph 00:09:55.120 --> 00:09:57.320 So let's invite Cheetah-Cub onstage. NOTE Paragraph 00:10:02.160 --> 00:10:05.816 So this is Peter Eckert, who does his PhD on this robot, 00:10:05.840 --> 00:10:07.896 and as you see, it's a cute little robot. 00:10:07.920 --> 00:10:09.176 It looks a bit like a toy, 00:10:09.200 --> 00:10:11.256 but it was really used as a scientific tool 00:10:11.280 --> 00:10:14.576 to investigate these properties of the legs of the cat. 00:10:14.600 --> 00:10:17.216 So you see, it's very compliant, very lightweight, 00:10:17.240 --> 00:10:18.496 and also very elastic, 00:10:18.520 --> 00:10:21.296 so you can easily press it down and it will not break. 00:10:21.320 --> 00:10:22.776 It will just jump, in fact. 00:10:22.800 --> 00:10:25.680 And this very elastic property is also very important. 00:10:27.160 --> 00:10:29.056 And you also see a bit these properties 00:10:29.080 --> 00:10:31.480 of these three segments of the leg as pantograph. NOTE Paragraph 00:10:32.280 --> 00:10:35.056 Now, what's interesting is that this quite dynamic gait 00:10:35.080 --> 00:10:36.976 is obtained purely in open loop, 00:10:37.000 --> 00:10:40.136 meaning no sensors, no complex feedback loops. 00:10:40.160 --> 00:10:42.576 And that's interesting, because it means 00:10:42.600 --> 00:10:46.616 that just the mechanics already stabilized this quite rapid gait, 00:10:46.640 --> 00:10:50.816 and that really good mechanics already basically simplify locomotion. 00:10:50.840 --> 00:10:54.136 To the extent that we can even disturb a bit locomotion, 00:10:54.160 --> 00:10:55.816 as you will see in the next video, 00:10:55.840 --> 00:10:59.736 where we can for instance do some exercise where we have the robot go down a step, 00:10:59.760 --> 00:11:01.376 and the robot will not fall over, 00:11:01.400 --> 00:11:02.976 which was a surprise for us. 00:11:03.000 --> 00:11:04.416 This is a small perturbation. 00:11:04.440 --> 00:11:06.856 I was expecting the robot to immediately fall over, 00:11:06.880 --> 00:11:09.316 because there are no sensors, no fast feedback loop. 00:11:09.340 --> 00:11:11.536 But no, just the mechanics stabilized the gait, 00:11:11.560 --> 00:11:13.136 and the robot doesn't fall over. 00:11:13.160 --> 00:11:16.296 Obviously, if you make the step bigger, and if you have obstacles, 00:11:16.320 --> 00:11:19.976 you need the full control loops and reflexes and everything. 00:11:20.000 --> 00:11:22.936 But what's important here is that just for small perturbation, 00:11:22.960 --> 00:11:24.456 the mechanics are right. 00:11:24.480 --> 00:11:26.576 And I think this is a very important message 00:11:26.600 --> 00:11:28.791 from biomechanics and robotics to neuroscience, 00:11:28.815 --> 00:11:33.495 saying don't underestimate to what extent the body already helps locomotion. NOTE Paragraph 00:11:35.440 --> 00:11:37.600 Now, how does this relate to human locomotion? 00:11:37.960 --> 00:11:41.600 Clearly, human locomotion is more complex than cat and salamander locomotion, 00:11:42.360 --> 00:11:45.496 but at the same time, the nervous system of humans is very similar 00:11:45.520 --> 00:11:47.096 to that of other vertebrates. 00:11:47.120 --> 00:11:48.576 And especially the spinal cord 00:11:48.600 --> 00:11:51.240 is also the key controller for locomotion in humans. 00:11:51.760 --> 00:11:54.176 That's why, if there's a lesion of the spinal cord, 00:11:54.200 --> 00:11:55.696 this has dramatic effects. 00:11:55.720 --> 00:11:58.496 The person can become paraplegic or tetraplegic. 00:11:58.520 --> 00:12:00.896 This is because the brain loses this communication 00:12:00.920 --> 00:12:02.176 with the spinal cord. 00:12:02.200 --> 00:12:04.416 Especially, it loses this descending modulation 00:12:04.440 --> 00:12:06.360 to initiate and modulate locomotion. 00:12:07.640 --> 00:12:09.336 So a big goal of neuroprosthetics 00:12:09.360 --> 00:12:11.736 is to be able to reactivate that communication 00:12:11.760 --> 00:12:14.200 using electrical or chemical stimulations. 00:12:14.840 --> 00:12:17.776 And there are several teams in the world that do exactly that, 00:12:17.800 --> 00:12:19.016 especially at EPFL. 00:12:19.040 --> 00:12:21.536 My colleagues Grégoire Courtine and Silvestro Micera, 00:12:21.560 --> 00:12:22.800 with whom I collaborate. NOTE Paragraph 00:12:23.960 --> 00:12:27.056 But to do this properly, it's very important to understand 00:12:27.080 --> 00:12:28.816 how the spinal cord works, 00:12:28.840 --> 00:12:30.536 how it interacts with the body, 00:12:30.560 --> 00:12:33.040 and how the brain communicates with the spinal cord. 00:12:33.800 --> 00:12:36.696 This is where the robots and models that I've presented today 00:12:36.720 --> 00:12:38.616 will hopefully play a key role 00:12:38.640 --> 00:12:41.296 towards these very important goals. NOTE Paragraph 00:12:41.320 --> 00:12:42.536 Thank you. NOTE Paragraph 00:12:42.560 --> 00:12:47.120 (Applause) NOTE Paragraph 00:12:52.100 --> 00:12:54.736 Bruno Giussani: Auke, I've seen in your lab other robots 00:12:54.760 --> 00:12:57.216 that do things like swim in pollution 00:12:57.240 --> 00:12:59.696 and measure the pollution while they swim. 00:12:59.720 --> 00:13:00.936 But for this one, 00:13:00.960 --> 00:13:04.440 you mentioned in your talk, like a side project, 00:13:05.640 --> 00:13:06.856 search and rescue, 00:13:06.880 --> 00:13:09.056 and it does have a camera on its nose. NOTE Paragraph 00:13:09.080 --> 00:13:11.576 Auke Ijspeert: Absolutely. So the robot -- 00:13:11.600 --> 00:13:13.029 We have some spin-off projects 00:13:13.053 --> 00:13:16.496 where we would like to use the robots to do search and rescue inspection, 00:13:16.520 --> 00:13:18.096 so this robot is now seeing you. 00:13:18.120 --> 00:13:21.296 And the big dream is to, if you have a difficult situation 00:13:21.320 --> 00:13:24.936 like a collapsed building or a building that is flooded, 00:13:24.960 --> 00:13:28.296 and this is very dangerous for a rescue team or even rescue dogs, 00:13:28.320 --> 00:13:31.216 why not send in a robot that can crawl around, swim, walk, 00:13:31.240 --> 00:13:34.416 with a camera onboard to do inspection and identify survivors 00:13:34.440 --> 00:13:37.216 and possibly create a communication link with the survivor. NOTE Paragraph 00:13:37.240 --> 00:13:40.816 BG: Of course, assuming the survivors don't get scared by the shape of this. NOTE Paragraph 00:13:40.840 --> 00:13:44.136 AI: Yeah, we should probably change the appearance quite a bit, 00:13:44.160 --> 00:13:46.976 because here I guess a survivor might die of a heart attack 00:13:47.000 --> 00:13:49.536 just of being worried that this would feed on you. 00:13:49.560 --> 00:13:52.416 But by changing the appearance and it making it more robust, 00:13:52.440 --> 00:13:54.496 I'm sure we can make a good tool out of it. NOTE Paragraph 00:13:54.520 --> 00:13:56.806 BG: Thank you very much. Thank you and your team.