Ep.15 Dr. Pauline Pounds - Associate Professor in Mechatronics at UQ and Senior Academic at Robotics Design Lab

Welcome to the Future Nearing Podcast. My name is Philip Saric. My goal with this podcast is to highlight some of the incredible science, technology and innovation happening in Australia and some of the key individuals who are helping to build the future. If you are interested in supporting the future nearing podcast or simply want to keep up to date with the latest episodes being released then I highly recommend you subscribe to the Future Nearing Substacker Channel. The link to this will be in the podcast description. On today's episode of Future Nearing I'm joined by Dr. Pauline Pounds. Pauline is an associate professor in mechatronics at the University of Queensland in a senior academic at the robotics design lab which he helped to co-found in 2012. Pauline has previously contributed to some of the foundational work in developing multi-roader drones. Now she spends a lot of time working on various robotics and mechatronics research topics which include a variety of aerial vehicle technology, bipedal walking robots, stabilization and control systems, applying deeper reinforcement learning to robotic systems and fighting innovative ways to make robots cheaper, better and more accessible. In this episode we discuss Pauline's background and a journey into academia, her work at the UQ Robotics Design Lab, some of the cool technology and research she is involved in and her thoughts on the commercialization of university research. Pauline thank you so much for joining me on today's podcast. Thank you very much it's a pleasure to be here. Yeah I'm super keen to get stuck into this conversation. Can we just start off by giving our listeners a bit of a background and a bit of introduction as to who you are sort of what you've gone through in terms of your career and your journey so far, maybe going back all the way to uni sort of what you studied there and then some of the roles you've had throughout your career to where you've ended up now in your role at UQ. If you give us a bit of a break down there that would be awesome. Yeah absolutely. Well I might actually take the initiative and go back a little bit further than that because my journey really began when I was five years old. I was one of those weird kids who always knew what they wanted to do and I knew that I wanted to be one of those people who made complex mechanical systems. When I was really young I saw a weirdly not the bed spread that my parents had with all these pictures of steam locomotives on them and they were just the most entrancing mechanical systems I'd ever seen all these moving parts and what did they do when I lay there and tried to figure out how they worked and that desire to understand mechanical systems, complex engineering structures, stayed with me my whole life and as I grew up I kept sort of fostering that interest playing with Lego and and other sort of construction like toys taking every opportunity I could to build things or they're out of cardboard or lumps of wood and these kinds of things to build an experience what it was like to make something complex make something real and I really carried that with me my whole life I thought when I was 16 years old I might become an aerospace engineer because they had the most advanced the most you know exciting technology of the era and as time went on I realized that you've got to be very smart and good at mathematics to be an aerospace engineer so I thought maybe I'll settle on something more like mechatronics and mechatronics is actually suggested to me by a teacher in high school who realized that I liked mechanical systems but also had an interest in electrical systems and I said well I don't know which one to do and he said well mechatronics is a thing have he looked at that so I've sloshed around trying to find a mechatronics program when I was looking for universities to attend and instead I found this thing called systems engineering at the ANU now I'm a bit of a weird achievement freak I was always one of these gifted talented kids growing up and when you're gifted talented kid the thing that the parents and teachers drill into is that your intelligence and your high scores or gives you value as a human being and so I wanted to go to the program with the hardest entry requirements because that seemed like something that would challenge me and it would meant I was winning at university admissions I guess I don't quite know but ANU had a had a program with an entry score of 90.6 some of the most specialized stuff like aerospace engineering was a little bit higher at a university all the way up in Queensland I thought yeah it seems that seems a little bit too far to travel and you know I'd sort of given up on that so I thought okay systems engineering what systems engineering well it turns out that we didn't know what systems engineering was even when I joined the program we actually had mugs made up with a little picture of a guy shrugging and the the motto what is systems engineering question mark? oh turns out systems engineering is sort of everything it's mechanical and electrical and structural and controls and dynamics and statics and photonics and semiconductor physics and classical electronics and computer vision and robotics and business studies and economics and law and accounting and all manner of other things that I can't even remember what we studied but there was a point where ANU was just switching to a name degree program so you could get majors in things and if I'd stuck around an extra six months and taken a slightly different elective I could have triple majored at ANU so it was really quite broad and ecumenical as a degree and it turns out that actually suited me just fantastically in the rest of my career I don't think there's any parts of my systems engineering background that hasn't coming useful at some point so after graduating with first-class owners from ANU in 2002 I set out and thought I would find a job of some sort because it turned out that after having structured my entire life around going to university and having found university to be this almost a wonderland of acceptance and tolerance you know I was bullied a lot in school and suddenly here at the university I was accepted for who I was and valued for being smart and not in the sort of destructive way that parents and teachers will value you for being smart and telling you to achieve harder for the sake of I don't know living vicariously or something three-year achievements this time around I was doing it because I was fiercely and passionately in love with the study and after having left this place and realizing wait a minute I have to pay bills now this sucks I went looking for work because unlike all the other students who are going through university contemporaneously with me I hadn't applied for graduate recruitment programs I hadn't applied for anything else at all never worked a job it had simply been almost incomprehensible to me as like well I've got my engineering degree what else is there to do what else is worth doing you know work as an engineer yeah I mean I guess I had just hadn't occurred to me that that was the thing that I should do and so I suddenly found myself out in the cold because engineering was actually going through a downturn at the time it was very hard to get jobs even as a graduate with first-class honors from the top university in the country at the time you just there was nothing going so I did what any sensible person would do and I went back to the university and said my old thesis project supervisor hey I enjoyed the thesis project I did have you got anything going now my thesis project was on an early quadruder back before quadriders were even a thing and no no I knew what these things were and I enjoyed working on the project it was a challenging thing we all worked with some really cool tech that was loaned to us by CSIRO this inertial measurement system is the little IMU the brains of the robot used it as the inner ear of the robot to steer the robot as it flew and the equivalent component these days fits on your pinky fingernail but it was a little black box about the size of of a coat can in size it was seriously heavy and very expensive and I was terrified of breaking it but we managed to get some results that were interesting and wound up actually being published and went on to be very very impactful and so I said to my thesis supervisor hey Rob have you got anything going you know you're doing a PhD and it just so happened that he had applied for and been successful in getting a discovery project through the ARC and he needed a PhD student so he says yeah alright and took me on there and the sort of that was that was it that's how I got into my PhD program and going back to university and starting a PhD was a very different world to me because I had been studying at the university in a very sort of structured way where you go to class your taught material you learn the material and you do the problems you do the assignments you sit the exam and you get a mark indicating your virtue as a human. Now suddenly, being a researcher, I was given a problem, a hugely wide open problem, and told, "Go solve it." And they don't teach you everything you need to know to do a PhD in your undergrad. You have to learn that yourself. And so for the first time in my life, I had this wide open vista of learning that I had to pick up and start running with. I had to figure out the fundamentals of my discipline. And credits my PhD supervisor, that he's a very smart man, but he was not a practical robot assistant the same way as I would need to become. He couldn't tell me how to solve these problems. It really was just up to me. And that was a very changing experience. I learned so much in sitting down in the lab and just reading books, doing the math, doing the simulations, calculating what was happening and what was going wrong, comparing it against the experiments I'd set up and finding out where I'd made the mistake or where things didn't line up. And I was an only child within the lab for a very long time. There were no other PhD students contemporaneous with me working in mechatronics or in robotics in the lab at that point. And so I really did fend for myself. And it a weird sort of way though, it perhaps stunted my intellectual development by not having a broader exposure to other people's work. It did give me the incredible benefit of being hugely independent and internalizing the idea that no one can solve my problems but me. And I think that's a really valuable lesson for people to get in their lives, especially if they're working in research. Because if you're dealing with problems for whom the solution may be found through the council of others, then perhaps you're not actually engaging on problems that are hard enough. You want problems where no one knows the answer, where being fierce and independent in your research is a real virtue. I think it's a good way to start, it's a very difficult way to start, but a good way to start. Yeah, following on from my PhD this time round, I made sure I didn't repeat the mistakes of my undergrad. And before I finished my PhD, I had applied for and been accepted to a role in industry. And at the time I'd sort of decided, "Oh look, academia, it's not really applied, it's not really real engineering, it's just kind of flapping about spending money, trying to learn things, it all felt artificial writing publications and attending conferences." And I thought I wanted to do something real, I wanted to actually make a change in the world. And so I applied for a job at a company doing electronics hardware development. I had been doing a whole bunch of circuit design just sort of as a hobby during the final drags of my PhD, when I couldn't stand just writing. I needed to get my fingers into an engineering project still. So I came up with some little hobby projects just to keep myself going, keep my sanity from evaporating while writing and writing and revising. And I really enjoyed it. I really loved doing the circuit board design. That's a skill set that still is with me today and I take a great deal of pleasure in exercising it when I get the chance. So I applied for a job being an electronics research engineer at a company called Locato, which conveniently enough was based just across the road from the A&U campus. And so right at the end of my PhD, when I started working there, I still hadn't finished making my drone fly. My project had been building this complicated quad-roader system that was in many ways still and extremely advanced vehicle today really pushed the bounds of what was possible at the time and even now many of its features remain really quite impressive. But we didn't make it fly at that point in time. So I felt sort of miserable having submitted my thesis without a positive result to report. So I did what any sensible person would do. And during my lunch breaks working at Locato, I crept out, broke into the lab with the key I still had and kept working on the project. I actually kept working there for about a couple of months actually. I would sneak in during lunchtime and on weekends preparing my aircraft for the final couple of tests to get it across the line even though my thesis had already been submitted. And eventually one foggy Friday morning we took it out from the lab, we took it to a flying field. There's a building there now that it was an empty field where we could fly. I think they used it for soccer most of the other time. And we flew it and it worked magnificently. And we achieved world-class result using techniques that sort of set the path for future drones in that space and the publication from that went on to be quite successful as well. So going from this sort of freewheeling, highly independent world where I was sort of captain of my own ship to a degree, even what I would have liked to have had perhaps a second mate position before I wound up driving the vessel myself without help to having a position where I was part of a much larger machine even though Locato was a start-up and reasonably small at the time. It was very different but also almost a kind of welcome. It was nice to be told to do this, now do this, now do this, and then follow through with my skills and knowledge to make those instructions actually work. It's funny because as a graduate with a PhD in the making, I was basically treated as effectively a fresh graduate from undergrad. And they said, "Look, we're not using any of the advanced skills so we probably shouldn't need to pay for them." I seemed reasonable, after all, I wasn't into engineering for money, money doesn't motivate me, it's useful for paying for stuff but I'm not in for a buck, don't tell the university, they'll stop paying me. I do it for the love of my discipline. And I thought, "Oh yeah, well, I guess I'll be okay with that." So I started working and the very first four weeks of my work was just building block diagrams of what their hardware actually was. And this really greated at me because I'd just spent eight years of my life getting to be a highly qualified and educated engineer and I'm spending my time doing documentation. That was not the way to start, especially when I'd been told that deep down, I'd be doing some really exciting development in electronics hardware. And I sort of said to them, "Guys, I'm here to do hardware. Why am I doing block diagrams?" I'm like, "Oh, well, you're educating yourself in how a system works." Okay, fine, fine. So the next project they gave me after that was to integrate an antenna system with a GPS receiver because GPS systems have an internal timing circuit that allows them to transmit very, very accurate signals from an atomic clock and your GPS receiver will then give you a one pulse per second measurement in time. And our system also had an internal time base and we wanted to slew the clocks of our systems to line up the GPS system. Okay, that's great. And now CEO said, "Oh, look, do as well as you can see if you can get it from say 200 nanoseconds there, maybe down to 20 nanoseconds there." Okay, okay, okay. So we started working on it and I realized that actually my experience doing control systems engineering was really useful here, as well as a bit of number theory. I spent a week just thinking about the nature of numbers, which brought me back to the joys of what I'd done during my PhD where I could sit there and really think deep thoughts about how stuff worked. And yeah, I sat down and I wrote this pretty wild algorithm actually to do the problem and do the problem to a level that I think they had no idea was even possible. Because when I showed them the results, they said, "Well, did you get it down to 200 or 20?" I said, "Oh, I got it to 200 years. That's not that great." I said, "No, no, no, no, no, two. What do you mean two?" And I showed them that I'd actually gotten the system alignment precision down to two nanoseconds. Far, far better than what they thought was even possible just because I understood the underlying mechanics of the system from a numerical perspective. And I had my performance review the next month. And I said, "So how about arrays?" And they gave me arrays, commensurate with that, which my PhD would probably have asked to have been paid. But it really set a fire in me. I realized that the work that I was doing wasn't fulfilling my soul. And I got pretty depressed and a bit aggro. My friends found me very unpleasant to be around because every time I caught up for coffee, I'd complain about my job and wasn't happy with the work I was doing and didn't feel like I was going anywhere. And I sort of talked about the door closing, that the opportunity for having a career where you go and you actually do something that's really emotionally fulfilling seemed to be drifting away from me. And I realized I didn't just want a job engineering. I wanted a job building something amazing, flying things, building robots. I loved it. I missed it. I really missed it. Working for the company just didn't nourish me. And at some point I said, "Guys, I had to just can't stay." And so I switched apart time and started looking for a postdoc. And I thought, "I'm either making the best decision of my life or the worst decision of my life, but it's a decision I have to make." And I applied to probably 40, 50 postdoc positions around the world before I found something. And there's a bit of a funny story actually. I applied for a job at Harvard where they were building robot insects. And another girl at the time who ended up becoming very well known in the robotic circles as well, applied for a job at Yale doing grasping. And we both sort of felt that we were a shoe in for the jobs we applied for. And it turned out that in their infinite wisdoms, the people who were running those two projects happened to know each other and discussed. And it turned out that we were each saw the others applicant as being a better fit. And so I wound up getting the position at Yale. She wound up getting the position at Harvard. And it wasn't until years later when we met at a conference and we talked about what our postdocs were. We realized we both applied for the others job and being sort of shuffled around to get the positions that we got. So very strange. But did I work down the end? We both had careers afterwards. Yeah, my postdoc position was a real departure from the work that I had done before. I'd always been focusing on building aircraft, making them fly or solving dense technical problems in a very constrained sort of way. But my postdoc at Yale was based on building a drone system that wasn't about the drone itself, but was actually about having the drone do a task that was grasping, manipulating, picking up objects, carrying them while flying. And when I told people what my postdoc was about, they said, "Good luck, that's impossible." And people literally told me it couldn't be done. And I'm not wanting to take a problem like that lying down. So when I got to New Haven, Connecticut, and started working on the project, I put a lot of thought into how to solve an impossible problem. And I realized that the nature of the problem was not just a manipulation problem, not just an aircraft problem. It was a unique coupled problem. And if you thought about it in a particular way, you could actually make the two problems solve the other by creating a coupled system, which was intrinsically stable, and that the manipulator would stabilize the aircraft in flight rather than destabilize it. And I found that a really exciting way of using my intellect to solve a problem by reframing the nature of the problem and thinking about it differently. And yeah, it ended up actually creating a whole field of aerial manipulation. No one had done it before we did it in 2010. We picked up objects with a hovering helicopter with a little flexible grip underneath. And if you look at all the stuff being done in Google Delivery Drones and Amazon Delivery Drones, they sort of traced their way back to the work from that era. And I'm pleased to report that the very first object picked up and transported by a helicopter outside, but robot helicopter, was actually a can of Guinness. And I have a photograph of it somewhere, it's great. It's great. Candy Guinness. And yeah, so it was a difficult time doing my postdoc. I didn't get along well with my postdoc supervisor. The first year was pretty good, but we put all this time and effort into developing the system. And we published our early results and sent it to RSS, Robotic Science and Systems, which is a real premier robotics conference. And we were rejected. And the editor's notes reported that they considered our work a gimmick. Literally their word was a gimmick. And I really want to look that person up again and say, hey, we created a whole new field of aerial manipulation. Who's a gimmick now? But no, I don't have the time to be petty, even if it would have used me to do so. But that made a sort of a turning point in our relationship, where I went from being a person who felt like they had a whole bunch of support to suddenly thrust out of out of favor and felt real pressure to deliver results and perhaps move on. And I ended up having a dispute that was mediated through the postdoctoral affairs office at Yale. They ended up supporting me, and that was great. And I wound up leaving under acceptable terms, I guess I would say. But it's really had a big impact on me in terms of myself confidence and the amount of stress I was at under. But the important thing is we achieved our goals for the project. And it really did set me up for my future career. And even though I didn't have a great relationship with my supervisor at the time, it turns out he actually wrote a glowing letter of recommendation for me, which helped me get my faculty position at UQ. So upon learning that it helped mend fences and perhaps bury the ax, I guess, between my supervisor and myself. Yeah, and that brings me up to about 2012, which is when I started my position at the University of Queensland. And it's been a journey, I've got to say, getting here. Yeah, absolutely. I mean, that's a pretty incredible story, even just the fact that, you know, when you were five years old, you already knew what you wanted to do with your life and sort of just having that alignment from such an early age, I think. I like to joke that all the other kids wanted to be astronauts. I wanted to make the rocket that those poor bastards would have to fly in. Yes, absolutely, absolutely. And even noting some of the differences you mentioned with your work in industry versus academia, you know, having more structure in industry, but also less freedom. Lessibility to exercise, I guess, your creativity. Yeah, it's very, very interesting. And I guess, yeah, maybe some people would prefer being in research and having that opportunity to, you know, really build some novel stuff and, you know, exercise their creativity as well. Well, there's different sorts for different types of work. I've met some students who are absolutely fantastic at just doing the mechanical step-by-step process of the engineering career that they follow the process from beginning to end. And they'd be absolutely brilliant working in industry in that kind of highly constrained environment. And there are people like me who are sort of wildly creative and a little bit anti-authoritarian who just chafen and wilt if put into that kind of environment instead. I think you've got to find the sort of career that matches your own internal speed, your own impedances, we would say, in the art. Yeah, absolutely. Yeah, I'm keen to hear more about your current role at UQ and within the robotics design lab. Correct me if I'm wrong. Will you, one of the people who co-founded or actually created that? Yeah, I was one of three co-founders. So myself, Hannah Kuhniowati and Syria Singh kicked it off in 2012. Syria was sort of the senior partner of that trio he arrived slightly earlier in 2011. And together we came up with the name because it sort of captured all the things we were doing and really had this beautiful ecumenical approach towards robotics design. In designing compasses everything, but this idea that both the robot machine and the robot algorithms could be designed. But most importantly, they could be designed together simultaneously to create a system that was really a gist out, really more than the sum of its parts. I think it really captured a philosophy that appealed to me a great deal. It had always been very interested in passive stability, passive mechanical design and doing things using just the electro mechanics rather than relying on raw processing power. And I sort of got to the point where I shied away from doing computer science a little bit. But the mechatronics, electro mechanics really spoke to me. And I know it's a weak part of me. I'm not as good at coding as perhaps I should be. But it did steer a lot of my work in a way that turned out to be really interesting and quite fertile. tile as well. Yeah, absolutely. So let's get into some of the work you're currently involved in some of the research projects and topics you're currently working on at the robotics design lab. Can we get into that? Oh, absolutely. In fact, I can give you a bit of a travel log about some of the things I've worked on over the past 10 years. And it's going on about 11 years since I started here. And there's been a lot of water under the bridge, but I'll try to give you a bit of a summary. So when I started at UQ, we had a little puddle of funding, no actual lab space, and an earnest eager desire to get up and running. And I asked myself, what can I do with no money, with no resources, and no lab space? And so I said, well, okay, what's the cheapest possible robot I could build? And I realized that in drones, the cheapest type of drone, the cheapest unmanned aerial vehicle was actually a paper plane. And we don't think about it as being an unmanned aerial vehicle, but it's a fully passive glider. It uses the mechanical structure of the folded aircraft itself to stabilize in roll, stabilize in pitch. And if you throw it, it'll glide for a long distance. And I realized, well, why don't I integrate that with an existing flight control topology and see if I can build a drone based on paper planes? And I came up with a neat design that leveraged existing paper airplane sort of forms and did a mechanical analysis, a structural analysis, an economic analysis on why paper planes were not a silly idea. Came up with a simple circuit board that you could stick onto it and could control the orientation of the drone in flight. And wound up publishing it the same year. And yeah, it was a bit wacky, but it was a way to get started. And I realized, well, the paper may not be the best material to use. What other materials would work? And I realized that flexible circuit boards, the kind of circuit boards that are inside mobile phones and computers and cameras that sort of can bend and fold and flex and be formed into shapes, could equally become aerodynamic structures. And I took that and made a whole bunch of sort of aircraft out of them. I made a glider very much like a paper plane just so that people could get the idea that flat circuits could form aerodynamic structures. And then I took a sort of a Samarice shaped like a maple seed, a little helicopter seeds that you see falling from trees. And I built a circuit board, which was sort of a Samarice shaped, but tuned to have a sensor package on board with a tropic circuit monitoring system like what you'd have tracking the weather. So temperature, humidity, pressure, ambient light, and a radio transmitter on the bottom. So it could send those values back to a base station. And the idea is that you could use them for bushfire sensing, almost like a fire alarm for the bush. And they worked very, very well. We ended up doing a couple of deployments. We built a drone vending machine that vends drones and flew around CSIRO's test area to plot where they landed. It's a fun project and a cool result. And that work was really impactful and wound up actually being asked if we could put it on display in the science museum UK just a couple of doors down from where they had Babbage's analytical engine on display. That was pretty cool to go visit them and see the thing there. At around the same time I started a project developing a new type of quad rotor. Now quad rotors as we know and love them, they have four rotors. I mean quad means four and rotor means spinny thing. So four spinny things. And as we know them today they're all roughly the same size rotor doing roughly the same job each. And I thought well look this isn't as efficient as it could be. Small rotors have to work very hard to provide the same amount of thrust as a larger rotor of the helicopter would because it can grab more air and push it down more gently. And I realized well why don't we build something more like that. The problem with helicopters though is that they've got these complex mechanical parts that control the orientation of the rotor in flight and direct the thrust as it maneuvers. And the rotor head the swash plate of a helicopter is notorious for requiring great deals of maintenance, requiring a lot of moving parts and generally being a nightmare deal with. So I thought well what if we combine the best features of a helicopter's efficiency with the huge simplicity of a quad rotor. I mean quad rotors are great. They're only four moving parts. And I said well okay let's have one big rotor to do the thrust and we'll have three little rotors to do the orientation control kind of like the legs of a tripod. You want to go forward, you tilt up the one at the back, you want to go sideways, you tilt one of the side ones and the direct she want to go. Now that just leaves changing your heading controlling your as it's called. And I realized that the main rotor of the helicopter needs counter rotation as well which is why a helicopter has a tail rotor to compensate for the drag torque of the main rotor. And so by tilting all of our three little rotors just a small amount to the side they could provide enough side force to all act like the tail rotor working together. And by changing the amount of thrust between the main rotor and those little rotors you could get a heading control for free. And it wound up being 15% more efficient than an equally weighted, equally sized, equal battery, equal flight control demonstrator. We actually built a sort of salt and pepper good and evil twin quadrotors to test with. The one without rotor design and one with the original classic quad rotor design. And we saw that 15% performance improvement. And the neat thing is that was without any particular optimization for a better quad rotor design. So by going to a larger rotor you also get aerodynamic efficiency improvements because the rotors larger the air acts more efficiently because you have a more developed flow. And if you tune it and optimize it you can reduce the amount of power required even more. And then because you've got one less rotor arm to deal with you have less mechanical structure. And then if you pack everything in more efficiently you can reduce the amount of cross sectional area that gets in the way of the rotor thrust. And then because your little maneuvering rotors don't have to provide as much thrust as a conventional quadrotors rotor which is doing all the lifting you can make its arms more narrow and therefore lighter and therefore have less obstruction again. And it becomes this virtuous cycle where eventually you can change the battery chemistry from being a high power lithium polymer cell to being a lower power but higher energy density lithium ion cell battery. And so it gets better and better and better and then you have the step change in energetic efficiency. And we go from having an aircraft which can fly for about 30 minutes to having an aircraft the same size with the same carrying capacity that can fly for 90 minutes. And we built this platform that was just absolutely revolutionary in what it can do. We went through the process of trying to commercialize that one and that's a bit of a tail unto itself which we might get into later on I suspect. Yeah so after that I had a project we were trying to sense the environment around a helicopter while it was flying around a drone. And we developed a type of thin whisker which could sense environmental flow since the aerodynamics of structures around it. But we actually got there by a an interesting process where we started out by trying to build a contact probe for aircraft that were contacting objects while hovering as you recall I was interested in building aircraft that could touch objects pick them up manipulate them. And getting accurate force measurements seemed like a useful thing to be able to do. So we built a special type of load cell using some technology that we shamelessly ripped off from Rob Howes group at Harvard University. And we built it into a long slender stick which allowed our hovering aircraft to poke objects and get measurements. And that worked reasonably well. We got some interesting measurements out. But then we thought why don't we make it smaller? And so by scaling it down we built a whisker which was almost diaphanus in its thickness. And we thought we could use this on conventional robots to measure proximity to people and objects and actually stop the robot from making movements which might be dangerous and hurt someone. And we realized that if we put it on a drone as it flew we could actually measure the fluids around it and identify whether the aircraft was drifting whether it was close to an object or not. And we carried on from that work and we built a sensor we could place between the aircraft and the rotor itself and measure the amount of force and thrust being developed by the rotor during flight. And we were able to show that if you're drifting sideways with the drone and actually applies an amplified force at the rotor which we could measure and compensate for our drift using that technology. And for the work that I did on these projects I got the 2020 Bata-Ham metal. I was really good work and quite impactful as well. So felt very humbled to be recognized for doing such important stuff and valued by other engineers across the country. It's almost a weird sort of dream to have wanted to be an engineer or your life and suddenly have a national engineering science body say you've done good. It's a real change in how I thought about my work. I'd always sort of done work because I thought it was important. I thought it was interesting, I hadn't thought that what I was doing was sort of something anyone would write home about. But no, it really gave me a boosting confidence. Yeah. No, that's incredible. You've worked some really interesting technology. What excites you the most right now in terms of some of the technology you're currently working on? Yeah, well, interestingly enough, the thing that's really of a star in my eye at the moment is completely different and yet almost entirely the same. So one of the things I've been getting into is walking robots. So bipedal locomotion. And you may be familiar with the really impressive work being done by Boston Dynamics, things like Atlas and Big Dog and Spot Mini. They're all these dynamical bipeds which balance by changing the forces between the robot and the ground in order to steer the body of the robot in the direction it wants to go and maintain upright stance. And as awesome as they are and as cool as the videos attached to the map, they are nevertheless extremely expensive. We're talking, you know, hundreds of thousands of dollars, if not millions of dollars, for example. And that's great, but it does mean that it's a bit beyond the access of people wanting to do things like package delivery or compound maintenance or inspection or security tasks. You just can't spend a robot that that's that expensive to do those those roles. And so I thought to myself, how can I make it cheaper? I've always had an interest in walking robots. I always thought they were super cool. In fact, flying robots were almost kind of a step down for me when I got into robotics to begin with because I really want to do walking robots. And it occurred to me that sitting down thinking about the nature of the problem, the problem with a walking robot is that they fall down. If walking robots didn't fall down, they wouldn't be difficult because then it's just placing one foot in front of the other and well, we know how to do that. Realizing that this is a stability problem long before it's a locomotion control problem prompted me to think, well, how to drone stay upright? Well, drones use rotors to develop torques that allow an onboard flight controller to steer the robot to maintain its upright stance. Well, why can't I do something similar for a walking robot? And in fact, in years since I started this project, other groups have tried building walking robots that use rotors to provide stability. It's a really cool solution. But I thought rather than using rotors where you use aerodynamic torque, why don't we use something like a flywheel? And you may be familiar with other robotics projects where people have used flywheels to stabilize them. And that's great. But those flywheels saturate really quickly. We needed something that allowed us to have a bit of a performance amplifier amplification there. And so I hit on an approach that was inspired by helicopter rotors because the gyroscopics of the helicopter rotor are quite important. And I thought, oh, I don't use gyroscopic procession. And it turns out that this technique is a known thing. It's called a control moment gyro. And so I took a walking robot and stuck a control moment gyro to it. And my simulations showed that it should actually work. The mathematical models said that I should be able to build a platform that weighs about a kilo, maybe a kilo and a half, and use sensible size actuators and sensible size motors and rotors. And so I started designing one. I had students at that point and thought, oh, this would be a cool project with students on. I don't have research funding for it. Let's just go it. Let's see how it goes. And we worked on it for about four years, building hardware and then tuning controllers and integrating things before we had it actually balancing properly. 2018, we had it balance and actually at my wedding honeymoon, we went to New Zealand and it happened to be a conference there at the time that we'd submitted to. And we presented this work at the conference and we ended up getting a best paper award. It's a really cool end to a honeymoon. But my husband would say that's about typical for me working during my own honeymoon. Since then, we've gone through a couple of evolutions of the technology, we have it walking. You can drive it around like remote control car, albeit very slowly at this point. The next version will be much more capable. But the real killer app here is not a consequence of the capabilities of the robot. It's just how mind-bogglingly cheap it is. Because we make the flywheel do all the difficult control tasks, we can make the system work with much more low cost actuators, the motors and things are so much simpler. The orientation control is literally just a drone flight control slightly repurposed. And as a result, instead of being millions and millions of dollars, it's just $10,000 or something like that for a human size robot. And working examples, even down to $2,000 for our small chicken size robots. So I'm really excited by them. We're right at the point where we'll be seeing a stepping capabilities pardon the pun. Yeah, that's super cool. I know there's a bunch of companies now working on this idea of general purpose, humanoid, bipedal type robots. And I guess this control and stability whilst walking is an important facet of that. And cheapness, because if no one is going to afford them, what's the point? Yeah, exactly. So it just makes me wonder, how far away do you think we are from having these humanoid robots walking around in their houses helping us with tasks? Oh, in our houses. Oh, look, I'd give it another 20, 30, 40 years, something like that. It's all about maturing the technology and getting it to a point where the cost of mass manufacture makes it affordable for people to have. I mean, the first ones are going to cost as much as a sports car, but eventually the cost will come down to where people can afford individual ones. The question is, what's going to make them actually useful to individuals? And I suspect that for a while they're not going to be doing the sort of robot butler task. I suspect they're going to be most useful, almost as sorry, humans for things like walking around conference centers or doing security, guard, patrol type things. It'll be little things in commercial space. And then eventually the AI performance will improve to the point where they don't need to be directly operated by human to do a task. And that's what's going to make it really democratized. I do see there'll be this sort of middle of the progression place where we'll have people in countries with lower costs of living, will tell the operate robots to live and work in the houses of wealthy people. It'll be the sort of weird dystopian future thing. It's going to be great. Yeah, I definitely look forward to that future. Things going to be amazing. Yeah, what sort of robotics and mechatronics technology are you most looking forward to working on in the near future? Well, once I've finished the walking robots because that's going to keep me occupied for a little while, at least I suspect. I'm interested in going big, I guess, going from small drone, small platforms that fly comfortably in my lab and allow me to explore the front ears of dynamical control to instead building things that are really tuned for working out in the real world. So things like flying cars, they used to be the sort of the realms of the crazy. They used to be the sort of, you know, almost career-killing thing to study. But now suddenly they're respectable again because we found the right acronym, Urban Air Remobility. I've had a whole bunch of interesting ideas in this space and I think I've got some cool ideas that allow us to break away from the real limitations that unduly are flying cars. Because if you think about it, a flying car is just a really terrible helicopter. And helicopters do a great job. That's why I built my new type of quadruder based around them. But helicopters have a whole bunch of limitations attached to them. And I think what industry needs to be doing is addressing the limitations of helicopters rather than trying to build something which is just too big, too large, too expensive and can't reasonably improve over the performance of things like existing general aviation. Hmm. Yeah, no, that sounds incredible. Yeah, I'm looking forward to flying cars in that future. That will be amazing. Yeah, we just have to get past I guess all the regulations. Oh, regulations as a whole kettle of fish unto itself. To me, the big problem is the engineering challenges. And I'm not even going to start looking at the regulatory stuff. There's a whole bunch of people who are excited and interested and skilled in that space who will be fighting that fight in parallel with me, I'm sure. By the time I've got something really cool that actually works, they're able to say, and here's how you fly it legally. Great. Yeah, absolutely. I'm keen to learn more about sort of this idea of commercialization. It's something that's come up a lot in this podcast and something I'm definitely personally very passionate about. I want to hear from you sort of what your experience has been, whether or not you've been able to successfully license or spin out any of the research and technology your work. looking on into products or companies. Tell us about that process, and if you've had any experience with that. Yeah, so I've actually got a fair amount of experience with it. It's one of those things that differentiates me from a lot of other roboticists and even academics in general. My philosophy is that there's no point in doing a whole bunch of really cool investigation and research and invention, and then just having it sit on a shelf in your lab. That doesn't do anything for anybody. And I'm pleased to report that I think without exception at the moment, every single one of my patents has been commercialized in some capacity, has been licensed or owned by a company. So I'm very keen to follow that process in the future and continue to take my ideas out there into the market. But it's a little bit gaulling as an academic because we don't get to be the ones who commercialize how work. We always have to partner with somebody else to do it. The university has some fairly strict limitations against us actually commercializing our own work ourselves. There are ways to do it, but it's pretty difficult. So the usual process that we follow is we have a clever idea and ideally a working prototype. We go to the commercialization body that we're working with. We tell them about it, they do the patenting or at least they go and talk to a committee about it and then the committee decides whether to patent it on on. If they agree to patent it, then they go out and they get the patent and then the clock starts ticking. And we've got to go out and find someone who will pick that up and run with it. Now ideally when you're doing your research, you already have a commercial partner in mind and best of all, you already have one lined up. One next life easier for everyone, but sometimes you've got to get out there and hit the streets and see if you can find someone who will give it life and keep the patent expenses paid for and hopefully do the commercialization and take it to market. And very rarely, sometimes industry will come to me and say, hey, we need a problem solved and I'll say I can solve your problem. Then they'll give us a bit of money. We do a bit of development, bit of research, come up with a solution, then we can patent it, they license the patent and then everyone's happy. And I've sort of gone through many different weird varieties of this. I've had instances where the research project has been driven entirely by industry for an industry problem and then they've gone out and they've actually started up a startup company turning that new technology into a product and fingers crossed. One of those will go well, we're just in the final stages of seeing if that can go to a market. I've had other ones where I've had a brilliant idea and I've supported it myself through my own research and sometimes even through my own funding to turn it into something that I think would be worth doing and I've found either students or other companies who are prepared to take on the tech startup company or else license it into an existing company. And actually before I went to University of Queensland, I was a co-founder of another startup company called Alaris and Alaris actually ended up becoming reasonably successful. We were a $90 million company at one point and unfortunately we weren't able to translate that into continued success and it wound up wrapping up in about 2017. But we built some amazing drones and I was able to work on it because it existed before my position at UQ and I had no real limitations but they were able to hire me through some of their funding, basically pay for my absence from the university to go and just sit down with an IP lawyer and pump out idea after idea after idea while I worked with them. They all ended up becoming part of the IP war chest of the company. It was pretty cool. Yeah, that's awesome. And do you think, do you have any ideas for how we can sort of improve commercialization? Do you think anything needs to change, I guess, for? Oh, absolutely. Absolutely. The way we do it at the moment is shockingly bad. And I'm not pointing the finger at UQ and UniQuest here specifically. I just mean Australia as a country has its heads screwed on wrong when it comes to this. First of all, we don't have a risk appetite. We don't have a culture of entrepreneurship in the same way that the Americans do. The Americans were willing to take a punt and just say, here, have some money, maybe it works out, maybe it doesn't. Let us know if you got something interesting. Whereas in Australia, we have this weird idea of hyper accountability where everything needs to be carefully measured and allocated and judged and made sure that everything is accounted for precisely and with precision. Nobody wants to fund anything unless there's 100% confidence in return on investment. And anyone who knows anything about business knows that that's a losing proposition at best. Because if you're not prepared to do something risky, you're not prepared to innovate. And if you're not innovating, if you're not taking a risk, then you're going to be beaten in the market by people who've already been there before you and who are doing something that's proven and well understood. You have to innovate. You have to take risk in order to get great reward. And our culture is just not set up for it. And as a consequence, it makes it very, very hard to do things like start-ups to create lean companies that are going to go out there and take a punt and maybe they succeed and maybe they don't. We need to get comfortable with the idea that one company in 20 will live and thrive and the rest will all die. We see that as being 19 failures. We really need to see that as one big success. If we don't start doing that, then we're never going to be able to leverage new ideas, new innovations. We keep expecting big companies to be innovative when in fact everything about the environment which they work is operating against innovation. They're already working well in the position that they're at. So why on earth should they spend a bunch of money to change up what's already a winning formula for them? You need young, small, hungry companies who are going to come along and try to eat their breakfast by being more agile, being more clever. And we need to give them the ability, the power, the resources to do that. We should be regularly expecting to churn through our companies as not just going to be the same old people, the same old companies who are running the roost. It's got to be a constant flow of new faces within our companies. If we don't have the opportunity for little companies to out-compete and displace the old, then really we're living with yesterday's approaches. And that's going to eventually come to tears if we don't have the ability to keep up internationally. We've got to be prepared to change. Yeah, absolutely. And do you find, in terms of collaborating with industry and industry coming in and wanting to solve a problem or build some new technology? Do they normally come in and fund early TRL research, or is it only they're coming in and wanting something that you're already currently working on and then funding that? How does that all work? In my experience, companies are generally not interested in doing blue sky stuff, at least not for the work that I do. As probably not true for the life sciences so much, I think they're much more switched on about it. But if a company is engaging with me, they're engaging with me because they want to have a problem solved. That means they want something that's ideally as close to TRL 9 as they can possibly get. And this sometimes leads to sort of a mismatch and expectation where they come to me to solve a problem. I develop a solution for the problem. And then they're all a bit surprised that I haven't turned it into something that shrink wrapped in a box that they can then put on shelves and sell to someone else. Science doesn't work that way. We're not commercial vendors. We work at a university, not a design consultancy. Even though we do a lot of the same things, our bread and butter's publications and new knowledge, learning new knowledge in the context of how to solve particular problems very valuable, but we're not actually paid or supported or really empowered to do commercial development. It gets way more expensive the further down that road you get, getting research engineers to develop products is just not good value for money. That makes sense. That makes sense. Awesome. I want to get into some advice for perhaps students or anyone interested in sort of mechatronics and robotics and potentially going down that route. What would you say are some important skill sets that people should know and learn if they want to get into mechatronics and robotics? And should they go to uni and study a particular thing? Should they do a PhD? Are they better off just learning at home and building some projects and looking online to see what they can build and learning that way? What are your thoughts there? Well, it depends on what you want to do. If you're interested in sort of working commercially as a roboticist, then there's some key skill sets, some core skill sets that you should have in order to do the job. So part of that is going to be the sorts of things that you learn hands on, but a lot of that's going to be things that you really only get through book learning. Now of course there's all sorts of types of book learning. Partly universities are an excellent way of doing it. But also there are things you can learn online through Khan Academy or through online courseware type stuff that basically gives you the same sort of experience. The big problem you have, especially in Australia, is that if you want to do the job, you're really going to need qualifications to compete with all the other people who would also like to be doing that job. So I really do encourage young people who are interested in doing robotics as a career to take the time to really understand what the nature of the employment environment they're looking at is, and certainly going to university is a great way to get a leg up there. You can do it by just being damn good at what you build and what you do, but then the onus of proof becomes much, much harder. And there's a bit of a glass ceiling when it comes to being a practical engineer who's learned the trade hands on rather than someone who's got a university degree or other accreditation. Because there are plenty of employers who will simply screen applicants by do you have a degree or not, and if the answer is no, they won't even consider you. And it sucks and it's not fair, it's not right. I think you can be a good engineer no matter how you learn how to do it. So long as you've got the skills you should be able to do the job, but the world we live in is one where employment is hard fought and getting a position doing robotics is particularly challenging to get noticed. Yeah, no, absolutely. And do you think there are any like course skills, you know, do they need to know PCB design or embedded systems or, you know, basics of computer programming or a bit of everything? What do you think people should really be? Yeah, so as a roboticist, your bread and butter generally is going to be, can you do software development in order to drive robot systems? Actually building the hardware for robots is quite unusual in Australia. If you're overseas, the answer may be somewhat different, but certainly in Australia, it tends to be software focused. So please do be good at doing coding, you learn languages like C or Python, those will give you a really good foundation C++ in particular. You should at least understand the mathematics of robotics, so things like kinematics, the study of movement and how you propagate forces and torques from the motion of actuators into the end effect of your robot or how you transform the motions. Of an aircraft into the trajectory that it takes and being able to do those things in reverse in calculating things like inverse kinematics and Jacobians and these sorts of things are, you know, again, key tools that students should know in looking for robotics background. Things like circuit board design is useful if you've got it, certainly things like understanding mechanical design is useful if you've, you know, if you've got the skill. But I think what will get you employed in Australia is mostly going to be your software and mathematical acumen. It's, I think a bit sad actually that we don't put more of an emphasis on co-design of robotic hardware with the software to drive it, but that tends not to be how the industry is moving because developing the mechanics of robots is very expensive and so it tends to be centralized and then the robots are replicated, developing software is cheap and quick and relatively easy to prototype. So long as you use one of the standard off the shelf robot platforms and so that tends to be where the skill sets are required. Certainly having a good grasp of physics, having a good mathematical background behind you, especially things like vector mathematics and differential calculus and differential geometry is especially useful if you've got that skill set as well. Awesome. That's some fantastic advice. Now I want to close off sort of my final question I love to ask my guests and the podcast is what is your vision of the future and how do you hope to help build that future? Oh gosh. So my vision of the future is this weird cross between the high fluton hyper optimistic George Jetson's future and the dark gritty miserable dystopian vision of Blade Runner. I think both will exist simultaneously and separately both at the same time. We all have flying cars roaring overhead while people live in work in a city that is full of robots and the robots won't replace us but they'll be part of what we do. There'll be tools that we employ, there'll be things we maintain, there'll be things that we sell, there'll be things that we buy, there'll be things that we program for fun. There'll be something that delivers our pizza, our boxes, our medicine and there'll be things where we study and we learn and we strive to deal with more and better and more capable robots as just part of our innate ambition. I think we'll see robots on the seas, we'll see robots on the roads, we'll see robots in the air and every capacity. You know, we'll have people complain the drones keep them up at night, you know, we'll have people complain that the delivery robots are clogging the sidewalk. We'll have people complain that you just can't get a good coffee from a robot barista, but you know what, it'll be okay. We all expect to see the dystopian reality where no one can work because robots do everything. But I don't think that's actually going to happen quite that way, not for a really long time. Instead we'll see the bounty of automation, make things more affordable. No one will be able to afford property anymore except for the property barrens you've snapped it all up. But that's okay because the people can't do anything about it because robot police forces will keep everyone in line. Hmm, no, maybe it's more optimistic than that. I don't know. Some very clever people will come up with ways in which robots can be more effective and they'll become new billionaires doing it. I don't think we've seen the first robot billionaire yet, but we will in time. It's pretty inevitable. You know, the person who writes the killer app for the robot that allows you to do all the washing up, they're not going to need to worry about money anymore. Yeah, absolutely. No, that's amazing. And yeah, it's a great vision. I think, yeah, I think robotics is an amazing field, an interesting field for people to get into. And I know mechatronics. It's still quite fresh and I think it's a really exciting area to get into so highly recommend for anyone looking to get into engineering to look at that space. I think it's very interesting. Yeah, I see the two are joined at the hip. I mean, mechatronics is mechanical electrical engineering and robotics adds the element of computer science to it as well. So there's something for everybody in robotics, depending on what you're interested in. Absolutely. Pauline, thank you so much for joining me on the podcast. Is there anywhere that our listeners can go? Maybe if they want to reach out to you or find out more about you? Absolutely. So all my details are on the University of Queensland website, so you can find me through there. Perfect. Thank you so much for joining me on the podcast. It's been great. You're very welcome. Thank you for having me. Thank you. Bye-bye. Cheers.