Bummer! This is just a preview. You need to be signed in with a Pro account to view the entire video.
Designing Robots for Real People40:06 with Dr. Ayanna Howard
The Robots are coming! The Robots are coming! The Robots are already… here. In recent months, there has been an upsurge in the attention given to robots and artificial intelligence, and their inevitable destruction of the human race if we’re not watchful. Whether your opinion sits on one side or the other, the fact remains; robots have already become a part of our society and, in some cases, an integral part. Ayanna will discuss the domain of intelligent robotics for real-world applications in health and in the home, and how these technologies can address real-life needs for improving our quality of life now and in the future.
[MUSIC] 0:00 [APPLAUSE] All right, so 0:04 when he said do you guys believe in robots, I saw like ten hands. 0:06 Okay, so I really need to do this again because I need to feel love in this room. 0:12 So, how many of you actually believe that if robots weren't real now, 0:15 they would be in say 20 years. 0:20 See, I got my answer. 0:23 So I'm gonna talk about robots. 0:26 I'm gonna start off by kinda showing you my biases. 0:28 So I have biases about robotics. 0:31 I think robotics are the coolest thing that ever existed because 0:33 I've been doing it for years. 0:36 I think there's nothing else in the world that's greater 0:38 than being a robotics engineer because I've been doing it for so many years. 0:41 So it started off. 0:45 Remember back in the day when everyone would ask you what do you wanna be when 0:46 you grow up? 0:50 What do you want to be when you grow up? 0:50 I was like I wanna be Wonder Woman. 0:52 I wanna be Princess Leah. 0:53 No, no, no, what do you really wanna be when you grow up? 0:55 Okay, I'm gonna be Buck Rodger's assistant, I'm going to be an astronaut. 0:59 No, no, no, what do you really wanna be? 1:03 And I saw this show called The Bionic Woman. 1:05 So this time, I wasn't gonna be a bionic woman, but 1:09 I wanted to build a bionic woman. 1:11 And everyone said yes. 1:12 This is something you can do. 1:14 Go ahead. 1:15 Go ahead. So, 1:16 at the time I thought I wanted to be a doctor. 1:16 Cuz, if anyone. 1:18 Okay, this is number one Bionic Woman, not the one they did as a remake. 1:19 Okay. So, if you remember the very first one. 1:23 It was all these doctors, and there were some engineers, but there were some 1:25 doctors and they would work on Bionic woman and build her up, and so I thought 1:28 I was gonna go to medical school in order to create and build a bionic woman. 1:33 So I was gonna go to medical school, I was gonna be this great doctor that did 1:38 all this surgery, and I took something called Biology. 1:42 And back then you actually had to dissect living creatures, they would come in, 1:45 the frogs were hopping, and they would actually cut them, 1:50 this was back in the day, and I think I was secretly a vegetarian then, 1:53 cuz I am now, and I was like okay I can't do this. 1:57 And if I can't cut a frog, how am I gonna actually cut the bad equipment. 2:00 So I actually had a teacher come and say hey, well, what about the parts? 2:04 The bionic parts, what about that? 2:08 I was like, what is that? 2:10 And she said engineering. 2:12 I was like okay, I'm gonna be an engineer. 2:13 I'm gonna go and I'm gonna be in engineering and I went to college, 2:15 and I said okay, I'm gonna be a robotics engineer. 2:19 Where's the curriculum, okay? 2:21 How do I sign up? 2:23 And I found out that robotics is a hybrid of everything. 2:24 So you learn mechanics, you learn electrical, you learn algorithms, 2:27 you learn computer science. 2:31 And so you are a Jack of all trade, and you don't go into depth in anything. 2:32 And so I learned how to build, design, program, and 2:36 work with people, how to do science, and all of this. 2:39 So what does that make me? 2:42 Makes me a really good robotics engineer and that's it. 2:44 That's why I think it's so cool. 2:46 So what do I do when I first started off, I decided, I'm graduating, 2:49 who does robotics in the world? 2:53 Okay, so I can either work for Ford, which really didn't sound attractive. 2:54 I could work for NASA, they do some cool stuff for robotics. 2:58 So I got hired at NASA and so I worked on Mars exploration, rovers, and designing 3:02 intelligence and trying to figure out what do scientists really want and think about. 3:06 It was beautiful, it was amazing. 3:11 I could that shirt that said, I am a rocket scientist. 3:12 Really? 3:15 Yes, I am a rocket scientist. 3:16 You better believe it. 3:18 Loved it, loved it, loved it. 3:19 And then we had an accident. 3:21 We lost a lot of our astronauts and I had to decide at that time, NASA decided that, 3:22 yeah, we really don't know what we wanna do. 3:27 Should we continue doing robotics? 3:30 Should we continue doing human space exploration? 3:31 Maybe we should go back to the moon, again. 3:34 Let's think about this and I decided that I want to do my own things. 3:37 I still haven't built my bionic woman. 3:41 So how can I do this? 3:43 It wasn't gonna be at NASA even though I was doing cool robotics stuff. 3:44 So, I went to Georgia Tech. 3:47 Amazing, I actually sent rovers to glaciers, to navigate, and 3:50 this is my stuff. 3:53 I then worked with kids, I worked with kids with disabilities, 3:55 designing therapy robots. 3:58 Getting closer to my biotic woman, designing exoskeletons and 3:59 things like that. 4:03 And so today what I'm gonna talk about is just the robotics as I see it. 4:04 And when I say designing robots for real people. 4:07 It really is designing the robots that will be out there serving and 4:09 will be in your hospitals and things like that. 4:14 And kind of give you a peek into what exists and what's coming. 4:16 So we have always, through out life, have a love/hate relationship with robots. 4:20 I mean, from the very beginning, so we love them, choppy, 4:25 he learns, he's like little kid, he crawls, and then we put a gun in his hand. 4:30 Love/hate. 4:34 We've had, if you remember back so, if you know this a lot of my images, sci-fi. 4:36 My bias so we have all these robots that exist. 4:40 Rosie is actually one of my favorite. 4:44 And she was sassy sometimes, so 4:46 she would talk back, which means she had intelligence. 4:47 But then she would also be very open, and actually cater to the family. 4:50 And so all of these robots we have this love/hate relationship. 4:54 And I've actually looked and said, why is it? 4:57 So it's not the same kind of relationship we necessarily have with, 4:59 say, automobiles. 5:02 It's not the kind of relationship that we have with, say, our computer. 5:04 We have some type of love/hate relationship with robotic systems. 5:08 Are robots gonna take over the world? 5:11 But you know what? 5:14 It'd be really nice if I had my robot chauffeur. 5:14 So this love/hate relationship has been in existence ever 5:16 since the beginning in terms of writing science fiction stories even until now. 5:20 And I say it's because robots, out of everything else, 5:24 the reason why we love robotics is that it functions like us. 5:28 We've created something that could be the next generation of humans. 5:33 And we hate them because they could be the next generation of humans. 5:36 And so, are we creating our destruction? 5:39 And so there was this binary kind of relationship that we have, 5:42 even in the same person. 5:45 And so why are we doing this? 5:47 And why are we designing? 5:49 We're gonna still continue designing robots. 5:50 That's gonna continue, but we're moving. 5:52 So when I first started robotics, I would say, I do robotics. 5:54 And the first thing people would say, you're taking the jobs away. 5:58 My grandfather used to work for 6:02 Ford and he got fired because of the robotic systems. 6:04 And that was when you had to explain technology. 6:07 Any new technology, you're gonna have jobs taken away. 6:09 But then you have new jobs, and you all have this kind of BS statement. 6:12 Now when I say I do robotics, they're like, when is my next robot aid? 6:16 When is my car gonna come? 6:20 When can I actually just sit in my car and go to work, and start tweeting and 6:22 not have to worry about anything else? 6:25 And so there's been this change about even perspective robotics. 6:27 The one kind of evil thing is there going to be a robot similarity 6:30 apocalypse or something. 6:35 But even then we have this, well we still should working for me but 6:36 we need to maybe have checks and balances. 6:41 Maybe we should have drones, maybe we shouldn't. 6:43 Maybe we should have some convention and laws of engagement. 6:46 And so again, love/hate. 6:48 But the conversation has changed so I'm gonna talk about robotics for 6:51 home and health. 6:55 And why I say home and health is because if you think about where you like your 6:57 robotic system to be, the closest thing in life is in your home. 7:01 And as for health, I mean we would all love to be able to live to 200 and for 7:05 me, if you can take my brain and my heart, and put it into a robotic contraption, 7:10 I can live to like 200 and see what the future was like, I would sign up. 7:14 And I think a lot of people would if they can guarantee that they are still human 7:18 in some aspect. 7:23 And so when I think about home and health, the things that we're looking at for 7:24 robotic systems, it's really to improve our quality of life to make sure that 7:27 our work life balance really is a true work life balance and 7:32 not a work balance and continue on. 7:36 And so I'm gonna show you a couple of robot systems that exist 7:38 in the entire domain of health as well as home and things like that. 7:42 And so, one of the things I wanna talk about the reason why we also look at 7:47 robotics for home and health is that everyone in this audience, 7:51 if you live long enough, are gonna grow up living with a disability. 7:55 So, what does this mean? 7:59 It means that the way we define disability is a change in your quality of life. 8:01 So, how many of you guys played sports when you were young? 8:06 Okay, a few. 8:10 How many of you still play sports? 8:11 Okay, less. 8:14 I love sports, but, that back doesn't feel like it used to. 8:15 When I try to run, not quite as quick as I used to be. 8:20 So imagine a world where, if I really enjoyed sports, 8:23 and I really enjoyed tennis, that, hey. 8:27 I can put on my little exoskeleton and guess what, my skill level is still there. 8:30 My body will not fail me. 8:34 And so, when we look at disability in the robotics world, 8:36 we really look at it as, if you have the need and you have a desire but 8:38 your body doesn't react or your mind doesn't quite get it. 8:43 It's not as quick little bit of memory loss. 8:46 If we can have that robotic assistant to help 8:49 we actually classify that as a type of disability, a change in function. 8:51 And so guarantee if you live long enough you're gonnao have a disability. 8:54 How many of you forget a little bit each day? 8:59 I had my keys. 9:02 I always put them in my pocket. 9:04 They're not there now. 9:05 When you were two, three, four, five, you're still developing. 9:07 You're still growing and so we change. 9:10 And so robotics, and disabilities, when we're designed for 9:12 individuals with disabilities, we're designing for the entire world. 9:15 And so that's really what, when you look at the robotics and 9:18 one other reason I'm talking about health, as well as the home. 9:21 So, going through some of the robotic process. 9:24 So, at the beginning, I'm gonna talk about platforms that exist, and 9:27 then at the end, I'll talk about my own platforms, and so 9:31 I'll put a little bit more love in the part at the end. 9:34 And so, if you think about robotic successes, in the robotics world, 9:37 what we define as a success is, 9:41 we have a system that has been used by individuals that didn't develop it. 9:43 And so if you think about it, it's beyond data testing. 9:48 It's actually to, say, you know, deployment with maybe a pilot group. 9:51 And so that's what we call a success, that you can actually take it, 9:54 you put it out in the real world and it works. 9:58 It only has to work with one, okay? 10:01 One and we're good. 10:03 It's a success because they're not the developer, which means that this 10:04 a different world, a different person who didn't expect, who didn't anticipate. 10:07 And so these are some of the robotic successes. 10:11 And we can continue on with that. 10:13 So one of the nice things in terms of rehab robotics 10:17 Is this whole new amazing progress in exoskeletons. 10:21 So, this is the ReWalk and so, the ReWalk, there's a number of companies, 10:27 some in Japan, ReWalk is out of the states. 10:31 And so, the ReWalk, this lady, 10:36 she's actually paralyzed from basically the waist down. 10:38 And the system straps on, she's sitting down, 10:43 she can actually stand and she can walk. 10:47 And so what's going on is the robot is coupled with her legs and 10:49 is actually moving and so its powering and actuating her legs. 10:54 They had an example of this if anyone's in to soccer. 10:57 So the world cup, they had an example where someone was wearing an exoskeleton 11:01 and actually kicked off the ball. 11:05 And so these kind of exoskeletons are in existence. 11:07 You can buy them, you can purchase them. 11:11 And so this is an example of a technology that started 11:13 because of providing individuals who are unable to walk, the ability to walk. 11:16 Recently, Japan just hired a bunch of these exoskeletons basically hired. 11:21 They bought a bunch of these exoskeletons to help their airport workers 11:25 pick up luggage. 11:29 They had a lot of workers that were getting a little bit older, 11:30 having difficulty picking up the luggage and 11:33 they said, what do we do in order to ensure the productivity of our workers? 11:35 So, they bought exoskeletons for their workers. 11:40 They're strapping them on and now you have these older guys coming and 11:42 lifting up luggage as if they were 20 again. 11:45 So this is an example of something that was designed for 11:47 the disability community, but now you see companies and 11:51 industry going, hey, it doesn't have to just be for disabilities. 11:54 This actually enhances the capability and allows them to lift and walk longer. 11:58 And so imagine, if you think about the future of Exoskeletons. 12:03 So imagine if we have a whole Olympics of people running. 12:08 And so really what it's about is your mind. 12:12 Can you design the best exoskeleton that can be at the 500 meter. 12:14 I think this would be a nice competition. 12:17 So now you can combine your mind as well as your physique 12:20 in terms of the next level. 12:23 Other areas that we're working on, Is prosthetics and orthotics. 12:26 And so, these are actually, I show these ones even though they're a little old, 12:30 because this is classically the first bionic man and bionic woman. 12:34 And so, I wasn't there, I wasn't first but I'll be there and 12:37 so, what happened is, both of these were amputees. 12:41 Claudia was actually military. 12:45 Jessie was an electrician. 12:48 He was up on a pole, got hit by lightning. 12:51 Fell off, lost both of his arms. 12:53 And so, what they do and what they design is, 12:55 they actually design a robotic arm and they took the pectoral muscles and 12:58 regrew the nerves and connected it to the robotic arm. 13:04 So, if you think about your hand out and you grab your fist, squeeze it or 13:06 you lift up your arm, all of these muscles are firing, and 13:10 they actually go up to the chest, and your brain is connected. 13:14 And so what they said is those muscles are still there, understanding how to grab, 13:16 even if your hand is, it's still there, and so can we basically take those muscles 13:21 and reroute them and regrow them with these nerves so 13:25 now you think about holding and grabbing and touching, and you still can. 13:28 And so this does exist. 13:33 So if you look at the future, what does the future look like? 13:36 Plastic surgery. 13:39 So you think of nobody wants a robotic prosthetic arm. 13:40 If I have no hand or I have no arm it makes sense. 13:44 But why would I get one willingly? 13:48 Well, how many people do enlargements, enhancements and do the facelifts. 13:50 These are surgery. 13:55 These are elective surgery that people do to enhance something. 13:56 And so imagine you're like, if I could actually have a connection to type my 14:00 code as fast as I think what my productivity level would be like? 14:04 Hey! 14:09 Sure! Give me your robotic hand and so 14:10 these are the things that people are actually thinking of. 14:13 Having these prosthetics to enhance your quality as you define your quality. 14:16 You know, I won't ever have those issues with my wrist or you know, 14:20 that little arthritis that sometimes bugs me. 14:24 have a robotic arm and everyone thinks it's cool, too? 14:27 So these are the kind of things of what the future looks like in terms of 14:30 something for 14:35 the disability community being pushed in terms of the typical mainstream community. 14:36 The other thing we have are methods for brain control. 14:43 This work, again, all of it is fascinating. 14:47 So when they start it off, this is a monkey what they did was, so 14:49 if you think about the prosthetic, it was connected to the pectoral muscle. 14:54 So it's not a direct connection, you still think it connects and 14:57 you have this interaction, this intent interaction. 15:00 But imagine you can actually grab directly from the brain. 15:04 So automatically. 15:07 So even before you consciously realize that you wanna movement. 15:09 Your arm is actually grabbing. 15:14 So this would be fascinating beyond many, many levels. 15:15 So it started off with the monkeys, to actually test it. 15:18 And so here, the monkey's actually feeding himself. 15:22 And so, I'm sure they starved him beforehand so 15:24 that he would actually want to eat. 15:27 But what's happening is that he's thinking about grabbing the food, and so 15:29 he is, okay, I see food, I need to move my arm, here's an arm. 15:34 That looks like something I'm controlling. 15:39 Okay I'm gonna have my arm and grab that piece of food. 15:41 Okay, let me bring it to my mouth. 15:43 And that's all through basically brain control. 15:44 Recently, they actually had, let me play this again, recently, 15:47 they actually had a number of trials. 15:51 Where they did the same thing. 15:55 So this lady, she was in a short trial about six months ago. 15:56 And they again connected her motor cortex, aspects in her brain, to this robotic arm. 16:01 And the first thing, when they first as you, if any of you ever do trials like for 16:07 drugs, some of the things they ask is what do you want to get out of this? 16:11 What is your goal? 16:15 Why are you doing this? 16:16 What is your ultimate thing? 16:17 And she had said all I want to be able to do is feed myself some chocolate. 16:18 That was her thing. 16:22 So if you think about chocolate, it's like yeah. 16:24 But look, chocolate you have to figure out how do you unwrap it? 16:26 It can melt. 16:30 How do you actually lift it to your mouth and eat it and then move it again? 16:32 And if it was really good, you'd want to do it again. 16:35 And so, very complicated motions versus just grabbing something and 16:37 sticking it in a hole. 16:41 And so that was her goal, was to actually feed herself a piece of chocolate. 16:43 And she did, so very successful. 16:47 Now, of course you see the plugs that are coming out of her head, so 16:49 this is not a doable thing in terms of going out into public. 16:52 But it's a step of understanding what that is, what that looks like, And now look 16:56 at the future they're now working so that you don't have to had basically invasive. 17:01 So in this one literally have open up your skull put in these nodes so 17:06 they can read the signal. 17:11 So imagine world now where instead of that you just put on little cap it 17:12 can be BCI cap you can read a really good BCI cap not the ones that exists now. 17:16 But really nice BCI cap, you read the signals, and now you control your arms, 17:21 you control your suit, you can now maybe actually think about military, 17:27 you can now fight with your robot partner just by thinking as if you're there. 17:31 So you're not there, but you're battling as if you were. 17:36 Just based on your mind control. 17:39 So direct connection. 17:40 So this is again the future. 17:42 So, in terms of my world, 17:45 where I work in is, I could it the target, the Walmart of robotics. 17:50 It's the robotics for everyday folks that just wanna basically have something. 17:54 And so the systems I showed you, they're still expensive so you can't like go up to 17:59 online, Amazon and say I want to get my robot prosthetic arm and it gets shipped. 18:04 That's not gonna happen. 18:09 So what I want to work on is robotics for everyday folks. 18:11 Low cost robotic systems that you can kind of forget. 18:15 So think about the Roomba Broom is actually a pretty affordable, 18:17 cheap robotic system. 18:21 Doesn't do a lot. 18:22 But, it vacuums halfway decently. 18:23 And so, those are the kinds of robotics that I want to do. 18:25 It's almost like you have the epitome of the iPhone. 18:28 I'm an Apple fan. 18:33 And then you have the smartphone that still works and 18:33 it's still functional and people are like, yeah, I got my free $29.99 phone. 18:36 And so I wanna do both, but it's still affordable for everyone. 18:40 So these are actually some of my robots. 18:45 I like to design robots for kids, that's my passion, and so some of my robots, 18:47 I have a robot that actually teaches you how to read and write, I have a robot 18:52 that does fetching, and so if you're sitting on the couch and you're like. 18:58 I really like my beer from the fridge and I don't want to miss this really good 19:02 commercial, you can send your robot to go fetch your drink. 19:06 And then i'll show you a video of my play robot. 19:12 And so thing that I have my robots do is a range of things. 19:16 And these are things that we sometimes take for granted, but 19:20 if you think about your quality of life, It changes. 19:24 So how many of you maybe, maybe not, you're sitting on a couch, 19:26 you want to change the channel, and you look around, and guess what? 19:31 The remote is right up over there, up on that counter. 19:34 You're like, I'm okay with this game that's on right now. 19:37 I'll change it a little bit later. 19:40 So, and this is us being healthy and just kind of eh, 19:42 we'll do it, my cognitive understanding of this is as good as I wanna get. 19:46 And so these are things that we think of as activities of daily living, 19:51 things that you take for granted. 19:55 You're getting into your car. 19:56 You're calling the Halo. 19:57 You're walking down the stairs. 19:59 You're feeding yourself. 20:01 You're dressing. 20:03 These are things we take for granted. 20:04 But for different individuals this might be hard. 20:06 So my ultimate goal for having a robotic assistant is a robot chef. 20:09 So I don't cook. 20:14 I'm married to a chef, there's a reason for that, I do not cook. 20:16 So imagine that I can dial in when I'm leaving work and say, you know what? 20:19 Today I really would like pasta. 20:24 Pasta with some pesto and maybe a little bit of tofu mixed in. 20:27 This will be really good. 20:32 Dial it in, and I get home and there it is. 20:33 All set out and it's actually good. 20:37 So, this is my ultimate desire of a robotic assistant in terms 20:39 of daily living. 20:44 Not quite there yet. 20:46 So what is it we have today? 20:47 So if you think about assistance in the home that exists for 20:49 individuals with disabilities as actually biological assistance. 20:53 And so there's a whole system of training monkey assistants to actually help. 20:56 They do things like go fetch your beer from the fridge, 21:01 actually open up cans, help you feed, and things like that. 21:06 There's actually some work where they're modeling with a big 21:09 robot the same kind of functions. 21:12 Imagine that you had difficulty shaving. 21:14 So I just say, imagine a shaver. 21:17 Imagine a robot holding it. 21:21 And the shaver gets closer, and you can't do anything except this. 21:23 Would you trust that robot? 21:28 And at some points, it's like no. 21:30 At others, it's yes. 21:32 And so some of those aspects they're actually working on. 21:34 So they actually have, there's always an engineer over the big red button, though. 21:37 Just so you know. 21:42 But they actually have robots doing things that you would typically say, 21:43 no I wouldn't. 21:47 But for a person who doesn't have that ability, this is amazing. 21:48 This is like, you know what, I will risk getting nicked a little bit. 21:52 In fact, they have diagrams of the first trial. 21:55 He had a few nicks. 21:57 I will risk that because now I don't have to ask my caretaker, I don't have 21:58 to ask my wife to come in and do something that's very basic such as shaving. 22:04 And so these things exist. 22:09 Where I go is, again, is for children. 22:12 These are my target, literally, Toys R Us robots for children. 22:16 So one of the things about robotics is that I don't even have to try. 22:21 So, if I show a robot to a child there is always already a natural, 22:23 oh, that's so cool. 22:29 What does it do? 22:30 I like that. 22:30 Let me play with it, let me interact. 22:31 And so there's always a natural affinity, and I get that from some adults too. 22:33 But kids will be engaged for longer periods of time. 22:37 The other thing is a robot is very patient, very repetitive. 22:39 So if I want to do something with a child such as rehab or therapy or 22:43 education, the robot will sit there and will not think, 22:46 oh my gosh I've told this kid this the fifth time, why don't they get it? 22:49 They will just go through the steps: one, two, three, four, five. 22:53 One, two, three, four, five. 22:57 So, very patient. 22:58 And so, this natural inclination. 23:00 And one of the nice things about a robot having this natural affinity 23:01 is that kids actually are trained to pay attention. 23:05 So you guys sitting here, listening, 23:10 watching my slides with the movies, this is actually a trained behavior. 23:12 When you were born, and there's a reason why, it's nature. 23:17 When you were born, you did not sit still. 23:20 I don't know if some of you might have been better than others, but 23:23 you would not sit still. 23:26 You would crawl, you would walk, you would explore. 23:27 And the reason is, is because it's your brain. 23:31 Your brain needs all of that input in order to engage with the world, 23:32 in order to learn, 23:36 i.e., you're grabbing things to taste it, not because you're just wanna taste stuff. 23:37 So it's like okay what is this? 23:42 Oh there's some texture to it. 23:43 Okay that doesn't really taste good, probably not going to eat that later on. 23:44 And so your understanding if I press that. 23:47 Oh look at that, what is that force? 23:49 Oh it hit me. 23:51 Okay, maybe I won't do that again. 23:52 And so you are actually exploring it, so the reason why kids are not initially 23:53 trained to sit there and pay attention is because their brain needs that input. 23:59 They need that constant input, it turns, look theres a light, there's a light. 24:03 There's another light, there's a person right there. 24:06 It needs that in order to do all that mapping, in order to do that learning. 24:08 And so we as adults then say come here, we're gonna do rehab, 24:11 we're gonna sit you, and it's pre-k and you guys are gonna sit there. 24:14 And we're gonna learn math, so robots are really good cuz they're like toys, and 24:17 so they can sit there and the child's actually engaged. 24:22 And as soon as the child changes, you can change it up. 24:25 And so robotics are ideally situated for this. 24:29 So what I'm gonna show you after this slide is some of my robots. 24:33 And talk a little bit about that. 24:36 And so when we design robots, we actually design them for play. 24:38 So, everything we have basically has an ulterior motive to get this child to do 24:41 something. 24:45 Whether it's again rehab, I just need you to do it over and over and over again. 24:46 Education, I need you to do this over and over again. 24:50 Anything that requires repetition. 24:53 But we do it in a playing activity. 24:55 And the way that we do it, if any of you actually have brothers or sisters or kids. 24:56 The way we do it is, one we actually had to, 25:01 you know how many parents put their YouTube videos of kids playing? 25:04 It's amazing. 25:07 So the first thing we actually had to do was understand the behavior of children. 25:08 What is the definition of play? 25:12 How do kids actually play? 25:13 Not only when they're by themselves. 25:15 But also with others. 25:17 Can we model that interaction? 25:18 When does a child become bored? 25:20 When does a child shift? 25:22 So if a child is playing with blocks, how long do they do that for and what 25:23 makes it so that they didn't switch and do something and, say, throw something? 25:27 And so we actually model all of this and 25:30 then we introduce the robot to then simulate that entire interaction. 25:32 And so we were pretty successful. 25:36 So, here's some of my robots. 25:38 And I'll go through all of them. 25:42 So, this is Darwin. 25:43 And in this case, Darwin is a interactive robot that helps kids with gait. 25:46 And so, imagine you have a child. 25:51 And typically, if you have issues with walking, or 25:53 if you have an injury, what they do is they make you walk. 25:57 So they make you walk, you have to turn around and walk again, 25:59 and they make you walk. 26:02 I mean, pretty boring, boring even for an adult, let alone a kid. 26:03 So imagine you have a robot that allows you to interact by playing soccer. 26:07 And so, and 26:13 I always call my robots, they're a little bit stupid, because what does this mean. 26:13 So I'm playing soccer with you, I'm kicking the ball back. 26:17 You're kicking. 26:20 I'm kicking it, you're kicking it, and all of a sudden, oh my gosh, 26:21 I make a mistake and I kind of kick the ball away from you. 26:23 Guess what the child does. 26:26 The child, like I'm not going to get that. 26:27 The child will actually turn and go over and kick the ball. 26:30 Kick the ball. 26:36 Kick the ball. Kick the ball. 26:37 Kick the ball. 26:37 And the robot, oh, made a mistake. 26:38 Oh, stupid me. 26:40 Kick the ball away, and the child will actually turn and move over. 26:41 And so what we've actually shown is that if you have a robot that's intelligent 26:46 enough to look stupid, the child will continuously engage with the robot. 26:50 And why is this? 26:56 Because it makes the child feel powerful, it empowers them, they know better, 26:57 they can teach this robot what they know, and so all of this basically empowers 27:00 a child to take control of their activities because the robot 27:05 is of course the most intelligent thing, and I can do better than the robot. 27:10 I'm awesome. 27:14 And so just this whole aspect of cognitive interaction or cognitive understanding and 27:15 pride actually makes the interaction that much longer. 27:20 The other robot we have which is, we call this the GT Avatar, is a Simon says robot. 27:24 So if you remember Simon Says, it was Simon says raise your hand. 27:32 And then you say, raise your hand, and he didn't say Simon says, so 27:36 you're out of the game. 27:39 So we put that on the robotic system, where we would have Simon says, 27:41 raise your hand. 27:45 And so the child would raise the hand, or move the fingers, or kick the leg. 27:46 And so we would have this movement. 27:51 And so one of the things we also found, and 27:53 all these are kind of lessons learned, is that you can actually 27:54 irritate a kid less than the parent, so what does this mean? 27:59 So if I'm interacting, and I say, Simon says lift up your arm, and 28:04 the kid does it and, what actually happens is robot says, I said lift up your arm. 28:08 Oh, come on, lift up your arm, we can do it, we can do it. 28:15 Oh. 28:18 Come on, lift up your arms. 28:19 So there is this engagement factor. 28:20 Eventually, the robot and the child connect and the child will follow. 28:22 In the meantime, the parent is like, are they done with the testing? 28:26 Can we get them out? 28:29 Because of this constant. 28:30 So a kid, in terms of repetition and cajoling them and 28:32 continually like, I'm gonna repeat myself over and over again until you do it. 28:35 A child will listen, not necessarily an adult. 28:39 So my favorite, I'm gonna get some sound to this, 28:42 So some of these are based on this whole aspect of Interaction. 28:56 And so, are they playing? 29:02 All right. Some of them are playing. 29:15 So, one of the things we found in terms of this engagement is that 29:17 there are certain things that people are naturally inclined to. 29:20 We can put them on a robotic system. 29:24 As an example, if I'm looking at you, and I go, okay I saw one person look. 29:27 So, normally, people will look. 29:35 I don't have to say anything. 29:37 Now if I do that and I say, look, you have more, but 29:38 I can do certain things with my body structure that actually evoke. 29:41 And so one of the things is that I can have big, 29:45 powerful, oh, she's trying to take command of the room. 29:48 Or I can say, [NOISE], which is a sad. 29:52 I don't even have to say anything, I don't even have to voice and so just through 29:55 body movement I can actually provide information about The interaction, and so 29:59 all of this has been studied by public speakers, how do you engage an audience. 30:03 Do you move your hands, do you keep them closed? 30:07 And so we take that interaction behavior and apply it to robotic systems. 30:10 And what comes out of that is a system where a child will engage and 30:14 the robot is basically, has this canned behavior. 30:19 It's not that we're learning it in real time, at least not the behaviors. 30:22 We learn this, we then model it, and then we interact with the child. 30:26 So I'm gonna show you what that looks like. 30:30 >> So, Darwin. 30:37 [SOUND] This is how you kill the pigs. 30:38 And if there's wood, 30:45 you try and get the wood. 30:48 Like that. 30:53 [SOUND] >> [SOUND] 30:55 >> [LAUGH] Okay, you try. 30:59 [SOUND] Thank you. 31:04 [SOUND] Your turn, Darwin. 31:08 [SOUND] Go ahead. 31:10 >> Okay. 31:13 >> [INAUDIBLE] [SOUND] 31:13 >> Okay, you didn't even have to see her 31:18 face and you know that he'd gone and screwed up something. 31:20 >> So try and- >> [INAUDIBLE] 31:23 >> Try and throw the bird higher. 31:25 Got it? >> So one of the things that we found out. 31:29 >> [SOUND] >> [LAUGH] Okay Darwin. 31:31 >> I'm gonna let it play, but turn off the sound some. 31:40 So, what's going on on the other side of the screen? 31:43 She's teaching Darwin how to play Angry Birds. 31:46 So we use the game, and if you think about Angry Birds. 31:48 You guys play Angry Birds? 31:51 Or used to. 31:54 Okay. So if you think about what you're doing, 31:55 is you touch the bird, pull, and you swipe. 31:57 Touch the bird, you pull, and you swipe. 32:00 You touch, I mean, that's it. 32:02 You touch, pull, swipe. 32:03 So very repetitive action. 32:05 The only difference is, how do you angle it. 32:07 A very repetitive action. 32:08 So he's sad. 32:11 He lost. 32:12 And so, what we found is that, so this is a learning algorithm so, 32:13 Darwin is learning, and this has to do with this whole aspect of, do you have 32:17 the robot learn optimally, or do you have it just learn little bit slowly? 32:22 And we found that, when a student, when a child, again, has that power and 32:27 knows more, they'll continue the engagement. 32:31 And also with the interaction. 32:35 So one of the things is that Darwin, he could figure it out. 32:36 Bits and bytes, Darwin can figure this out. 32:39 It's an angle, it's geometry. 32:42 Pretty simple. 32:44 But in terms of the child, the child will continue showing it. 32:45 And so, a little bit of tricks of the trade in terms of intelligence. 32:49 If I think about communication, so if I'm talking, da, da, da, da, da, 32:53 I'm talking, and then I stop, and I'm having a conversation. 32:57 I stop. 33:01 Okay, usually, someone will talk. 33:05 Okay? So, with Darwin, the way he understands or 33:07 the way he'll comment is, she talks, and some kids talk more than others. 33:11 And she'll talk, when she stops, it's assumed that she said something like 33:15 it's your turn or good job, or let's do it again. 33:20 And so what Darwin is actually listening for, doesn't understand language at all, 33:26 just understands that if I hear noise at a certain frequency level, that's speech. 33:30 And then if I don't hear something, 33:35 we're gonna assume that she said something like let's take a turn. 33:37 And so just something very simple as an algorithm. 33:40 And if you look at oh, this is my best friend, so intelligent. 33:43 I gotta teach him how to play Angry Birds. 33:47 Just by doing something very simple. 33:50 In terms of interaction, 33:52 the interaction which is sadness and happiness, very remote kind of, 33:53 if you think about the algorithm, if you lose, typically you're sad. 33:58 And so the algorithm is if you lose, or if you did worse than before, 34:02 you're probably not happy so, we're going have a sad. 34:06 And you can do a randomized sad because sad looks differently. 34:09 But if you win, or you did better, that's happy. 34:11 So, some very simple behaviors. 34:14 And yet, the robot seems highly, highly intelligent. 34:17 And then the last thing is this whole aspect of 34:20 putting in a little bit of behavior so it's not as smart. 34:23 Because again, you have all the data. 34:26 You can figure it out, the robot can figure it out. 34:27 But, as a child if I'm teaching someone and they get it and they get it and 34:29 then they don't do it. 34:34 What I want to do is, they're learning, okay let me show that to you again. 34:36 And then they get it, they get it, and then they don't do it. 34:40 I'm like, let's do this one more time. 34:42 And so what happens is the robot can figure it out. 34:44 But, just like putting in that little bit of I'm gonna be stupid right now for 34:48 this one turn, what we saw is that the child continues engaging and 34:52 engaging and engaging. 34:57 Whereas without the robot doing this, and 34:58 we actually did some tests of non-interactive or 35:00 a very intelligent robot and the kid after a while was like I'm done. 35:03 I'm done! 35:06 Robot can play by himself, I'm done. 35:08 I'm going home. 35:11 Again, having a robot that's just a little bit less intelligent, 35:12 even though it could be, is ideal for interacting with real people. 35:17 And so if you think about, okay, so I showed these robotic systems, 35:22 rewalk some of the systems that I do out of my lab in my company. 35:28 But how many of you have robots? 35:33 I remember there was Roomba. 35:35 Yeah, so it's not a lot. 35:37 So why is that? 35:38 Why don't you guys have your robot? 35:39 Why can't you go down to the local, you know, department store and 35:41 go to the electronics section and go, okay, 35:45 today I want my robot chef and tomorrow when I get a little bit more of 35:48 a paycheck, I'm gonna go buy my robot chauffeur or go online. 35:53 And so there are some issues. 35:57 There's still some issues. 35:59 So if you think about the technical issues, so one is, 36:00 as an engineer, computer scientist, I know how to talk to my robot. 36:03 I just program it. 36:09 What about the general public? 36:11 And typically, general public, not necessarily techie. 36:12 A lot of older adults, for 36:16 example, they don't embrace technology like we would want them. 36:18 Kids, kind of good at it. 36:21 So how do you communicate to your robotic system? 36:23 And one of the things is that people will adapt to technology. 36:26 So how many of you guys have an iPhone? 36:29 Okay. 36:31 Swiping pinching? 36:32 Do you know before iPhone one, what did you do with the phone? 36:35 Like Blackberry. 36:38 Exactly. 36:41 And so now in fact, if you notice, sometimes kids will go up to a terminal 36:41 that has a keyboard and they'll be like, oh god, I gotta press. 36:45 And so that is something that we have adapted to in terms of technology. 36:50 And so this communication, people will adapt within a certain domain, 36:54 in terms of communication. 36:57 So, what does that mean? 36:58 So if I say, you can communicate any way that you want to, but 36:59 this is your dictionary of 10 terms. 37:03 So you can say whatever you want, but if you really want your robot to do 37:05 something, these are the terms that you have to use. 37:09 People will adapt. 37:11 The other thing is, is how do you ensure the robot engages the person? 37:13 And so we found that some adults, you know they want a robot that is slightly. 37:17 Most kids want a robot that they can have empowerment. 37:22 But we have some kids who are really really stubborn. 37:26 And what does that mean? 37:29 They need a robot that yells at them like their teacher. 37:30 And so how do you recognize the personality traits of the individual that 37:33 you're working with, in order to change how the robot reacts? 37:37 Because again, these are robots for everyday people and 37:40 everyone has some slightly different nuances and characteristics. 37:43 And so how does a robot actually identify that 37:47 when they're interacting with a person? 37:49 Maybe it's you have to fill out a questionnaire that says, you know, 37:51 do you like people yelling at you? 37:54 Yes or no. 37:56 Yes, no. 37:56 Maybe that's the way we have to do it but this is still a challenge. 37:57 And then the other thing is, is how does the robot actually do a behavior? 38:00 What kind of things does a robot do in a safe manner? 38:04 So example, with the shaving, the way that they figured out the shaving, 38:07 like I said the first trial they had, and we had a little bit of nicks. 38:11 So what they actually did, how they solved that is that they would actually have 38:15 the robot get close enough and then what the person had to do was push down. 38:19 So that's how they solved it. 38:24 That's the kind of things about how do you have the robot provide enough feedback 38:26 in actuation and behavior, so that you change something that 38:30 could be a very unsafe kind of situation to something that's like, oh, ok, 38:33 I know the robot can't nick me because I have that last bit of control and 38:38 I can move my head a little bit, and then I can do the shaving. 38:41 And so thinking about robotic system and designing that 38:44 is really the challenge because everyone interacts with technology differently. 38:47 And so, the way that I interact with technology is definitely gonna be 38:51 different than someone else and someone else and someone else. 38:55 And so, how do you have the robot evolve to that? 38:58 And, of course, always it's about the money. 39:00 And so all of these robots, there are robots that are chauffeurs, 39:03 there are robot chefs that actually exist. 39:07 There's robot receptionists. 39:09 In fact, Japan just opened up a fully automated robot hotel. 39:10 Where you actually go in and you register, you get your key and 39:15 your bags get delivered and it's fully automated with these robotic systems. 39:19 So these things exist. 39:23 But they don't exist at the Target or the Walmart level, and so 39:25 that is really also a challenge is to get it from being very, 39:28 very expensive but existing all the way down to okay, 39:33 everyone can afford it and it's still functional, and it still makes sense. 39:37 And hopefully, it won't take that long. 39:42 And always one of the questions I'm always asked is, 39:45 based on your experience how long do you think it's gonna take til we get to X? 39:47 And my answer for the last 20 years has always been 10 years. 39:53 >> [LAUGH] >> And so today it's gonna take 10 years. 39:57 And I am done if you have any questions. 40:01 >> [APPLAUSE] 40:05
You need to sign up for Treehouse in order to download course files.Sign up