August 23rd, 2012

Transcript – Robotics festival

Robotics festival

Interview with Francesco Mondada, followed by a collection of short interviews conducted on the floor of the EPFL Robotics Festival

ROBOTS: Hi Fancesco, great to be talking with you after yet another successful robotics festival. We interviewed you last year about your excellent work in educational robotics. For those who do not remember, can you tell us what this festival robotics is?

Francesco Mondada: The robotics festival is an event where we have people with a passion for robotics sharing this passion with the public basically. So it is more than an event on robots, it is an event on people loving robots and sharing their expertise or their experience, and this is what the festival is about and what makes it successful. People really feel this passion and are enthusiastic about this.

ROBOTS: What happened during this year’s edition?

Francesco Mondada: This year we had, like every year, shows, exhibitions, workshops. The workshops are the most demanded activity because people really want to come and do something especially kids, especially their parents, we can say. This year we had the same mix, we had an addition, animation for everybody which was to build small robots with 10 bricks and this was very funny because the adults really were skeptical about this done by the kids. But the kids were enthusiastic and then at the end we had a really great event at all the level, 2000 places in the workshop and then more than 6000 places in the shows. We had the Air Jelly from FESTO this year, which was just wonderful. It is not a very complicated robot. I was surprised that my kids understood how it was working so there was nothing really incredibly complicated but just beautiful and impressive. So that is exactly what we look at for the festival, something which is great, impressive and people can understand, that is the best because we can really show that robotics is accessible and people can understand technology

ROBOTS: You mentioned the bricks. What are some of the other things kids are doing at these workshops?

Francesco Mondada: During the workshop kids can solder, can program, can build, there are plenty of activities, can dismantle systems, and this year I got some feedback of adults saying why is this limited to kids? I would like to do it myself. And I told them, just apply we do not check the age, so half of the people registering for the picobots, this world construction robots were adults.

ROBOTS: You had 15000 people this year. Why do you think this festival is so successful?

Francesco Mondada: I think the point number one is the passion of the people that are animating the festival, so this year we had nearly 600 people of staff to manage the festival and 350 are volunteers that are passionate about robotics. There are students who could be paid to help, but there are also many students that are really passionate about this activity. So I think the trick is to have people who really want to share their enthusiasm about robotics and technology in general.

ROBOTS: Do you have any fun examples of something that the public said or something that happened during this festival this year?

Francesco Mondada: Oh, I think that we had many, many, so more than every year, we have huge feedback by email and also during the festival. This year we doubled the space and we redesigned fully the registration system. Basically nobody had to wait for anything, so people were just enthusiastic to be able to go around freely. The only criticism I got was, there are too many things to see and we cannot see everything in one day. But I am satisfied with this type of criticism, it is not a problem. But it is becoming huge and we got a huge number of positive feedback because it was current, and I got also extremely good feedback from the exhibitors, the people making the show, etcetera, saying the public is very interested, it is silent during the shows, etcetera, so people were impressed by the atmosphere that was there, both from the side of the staff and the side of the visitors which was very good.

ROBOTS: Do kids know what robots are now? Do you still have to explain that to them?

Francesco Mondada: I think that you do not have only one explanation. What is important is that the kids get a little bit of understanding that first they have to know that they can understand, because even with the adults that we had a discussion with, even the adults say, in any case I cannot understand robotics so they even do not make the effort. I think that we have to pass the message that this is not so incredible, that people can understand what is behind it, at least part of it, and they can get an idea. The Air Jelly this year was not an incredibly complicated device. It was quite simple but quite efficiently impressive, so I think that this is the message we have to pass, and if the kids understand that they can dismantle something, understand what is inside, etcetera, then the goal is reached. We do not have to explain to them all the routes and all the mechanisms but we have to trigger this curiosity and destroy this fear that there is a lot of complex technology

ROBOTS: Do you think all these kids are going to become roboticists in 10 years are going to have a huge wave at EPFL?

Francesco Mondada: Frankly speaking, I do not care. As long as they keep this curiosity, we would have enough of them coming as roboticists, I think. But what is important is that they keep this curiosity. We have many of them which will not come on EPFL for sure. But there are plenty of them who are interested even to make other careers in the technical field, and this is important. So this year this was a little bit new. We had plenty of professional school who came presenting the school, etcetera. For those kids who want to do a career, not in academics or engineering, but more in a practical job related to technology, I think this is important that we are also at this level of support.

ROBOTS: What were some of the interactions with the NCCR with this year’s edition?

Francesco Mondada: The NCCR was very good in bringing the communities and the several labs that are involved, and we had more people from Zurich than usually, and then many tools that are developed within the NCCR, for instance for education, has been used in the workshops. We had many workshops organized by NCCR people and we use the festival as a test bench for some experience for the NCCR. For instance, the small robots, the picobots, we had 2 possible activities; one with a challenge and one with creative activity and we recorded every kid on each activity, what choice he made and we have statistical data on the job of the parents, on the age, etcetera, and we will now work with some people in sociology to analyze a little bit what is motivated for the kids in this type of activity and which kids choose what type of activity. So I think we will understand a little bit better how to attract kids and what type of profile we are attract with what type of activity, and this is something that we would bring within the NCCR and the NCCR will bring to the robotics community in general.

ROBOTS: What would you like to see for next year’s edition if there is anything more that you can do?

Francesco Mondada: This and the last year we pushed the participation of the teachers. So we try to involve the public school teacher within the festival to help them to use the tool that we use in the school. My dream is to see the fusion of the festival into the school more and more. So we got extremely good feedback this year. We are even now closed from school, we want to have a picot animation in our school, we want to have this and that. So my dream is really to be able to extend this local and big activity over all the year, and be able to spread this to all the people who do not come to the festival, because basically those who come to the festival are those who are already interested by robotics and technology, and if we can spread a little bit this activity also to other people that are not necessarily in the field I think would be great.

ROBOTS: How do you think that can happen? Is it about making the kid or is it about making the team of the team of people who go around?

Francesco Mondada: I think that there is a big problem with the teachers. They are not trained at all for this. They fear technology as an additional complex tool that they will not be able to manage or that kids manage better than them, this is also a big fear from the teachers, and we have to mange a way to help them to acquire these competences, to acquire confidence in the tools that we propose to them, and give them material to do their job in a more efficient way using technological tools.

ROBOTS: You have a lot of great projects in your lab. Can you tell us a bit about some of the more recent developments?

Francesco Mondada: We have plenty developments in the field of education, of course, related to the festival, related to the activity in the schools, but we also look in general to the use of robotics in their life, and we explore a way which is a little bit different from the classical approach, which is bringing robotics in everyday objects, and we are doing a set of experiments now also with kids, bringing a robot in their sleeping room and helping them to tidy up the room. So we are doing some experiments, and it seems very positive, at the level of acceptance from the kids and at the level of acceptance from the parents, and I hope that we will also bring a little bit new ideas into robotics in that way.

ROBOTS: I would have loved to have that. What does that robot look like?

Francesco Mondada: It looks like just a box. It is just an interactive box and the box is happy when something is thrown inside so kids understand very quickly and, if you put this facing the kids, after 5 – 10 minutes, all the toys are in the box, no problem.

ROBOTS: That is smart. Thanks Francesco

Francesco Mondada: Thank you to you.

Narrator (Jana): And now, let us go on a tour of this year’s festival to have a look around and hear first hand from some of the people on site with our reporter, Amanda

ROBOTS (Amanda): Hi Sven, how are you doing?


ROBOTS: Can you tell us what you are showing here?

INTERVIEWEE 1: Here we are showing basically a demo about robots that are following each other. So it is a simple demo that basically uses a simple interaction system to show something that is emerging so robots making long queues of robots.

ROBOTS: Looks really nice. So this is very interactive. I am sure this is an awesome project for the kids. What do you think they are learning from this?

INTERVIEWEE 1: They are learning to place their finger in front of the robot basically, but I hope they are learning a bit about how sensors work and how motors are actuated, depending on how the finger is placed in front of the robot.

ROBOTS: Right, very nice. There is a lot of work behind these small robots and often the ideas behind them are very visionary. What do you do to convey this concept to the public?

INTERVIEWEE 1: Here we keep it simple so we try to show like how simple programs, simple interactions can create emerging behaviors for collective robotics in our example. But we try to keep it as simple as we can to get core concepts.

ROBOTS: Great. Okay thank you very much then.

INTERVIEWEE 1: You are welcome.

ROBOTS: Now we are going to go see Adrian, see what he is doing. Hi Adrian.


ROBOTS: How are you doing?


ROBOTS: Can you tell us a little bit about what you are doing here?

INTERVIEWEE 2: Here I am trying to demonstrate an application of a robot that finds an odor source.

ROBOTS: So your odor source is actually ethanol, right?

INTERVIEWEE 2: Yes, which is quite funny to some of our owners especially when the robot comes towards them.

ROBOTS: And you have a fan there creating a plume so that the robot is actually trying to detect the plume and drive up the plume right?

INTERVIEWEE 2: Yes. In fact on the robot we have two sensors; one is for detecting the odor source, the odor itself and the other one is for detecting the wind direction. It needs these two information to implement a biologically inspired algorithm which is inspired by moths. The moth, in order to find a source, it will go in the direction of the wind while it senses the smell, and when it loses it starts spiraling until it finds it again, and then it just alternates between the two.

ROBOTS: This is actually a pretty impressive piece of work in terms of hardware and software. What are the future applications for this?

INTERVIEWEE 2: We might still be a bit far from applications. The closest applications that we could easily go towards are for industrial scenes where you have a pretty well defined environment but in the future the idea would be to use them even in rough terrain, for finding mines or for finding different leaks of toxic materials and so on. Applications are many, but we still need to work on the research.

ROBOTS: Right. Well thank you very much Adrian.

INTERVIEWEE 2: Thank you.

ROBOTS: Have a nice day… Hey Adrien [not the same one].


ROBOTS: How are you doing?

INTERVIEWEE 3: Very good.

ROBOTS: Why don’t you tell me a bit about what you are showing to the public today.

INTERVIEWEE 3: Sure. I am showing them this airborne robot. It is kind of a strange robot actually. They are always a bit surprised to see it. As you can see, it has plenty of protection around it. That is so that it can run into obstacles and even fall on the ground from 2 meters for example and have here some crashes.

ROBOTS: Right, looks very robust. So these are actually flexible bumpers right?

INTERVIEWEE 3: Definitely. These are bumpers all around it.

ROBOTS: Right, very nice… Hi Maja, how are you doing?


ROBOTS: This is an unusual piece of work here. Why don’t you tell us a bit about what you are doing

INTERVIEWEE 4: So this is a fixed wing, a flying robot which we use in outdoor environments, in a swarming project. So the project is about using flying robots in a search and rescue operation. In order to provide communication we [relay between] rescuers on the ground and to guide the dogs in a search and rescue, which is the first application of robot-animal interaction between dogs and flying robots.

ROBOTS: So this is during the survey and it communicates with the dogs.

INTERVIEWEE 4: Yes, and it works [? every time you can distribute it at the centralized link. ?]

ROBOTS: Very nice. What does it tell the dogs to do?

INTERVIEWEE 4: So basically it performs some kind of attractive trajectories and the dogs have to be trained to perceive the planes as helpers and they have to understand those trajectories and then they have to follow the plane.

ROBOTS: So there is a lot of training here involved as well.

INTERVIEWEE 4: Yes, this is a process we just started but we hope the dogs will obey us.

ROBOTS: Sounds like almost a new domain in robotics.

INTERVIEWEE 4: Yes, it is completely new.

ROBOTS: It is not human-robot interaction, interesting.

INTERVIEWEE 4: Animal-robot interaction because they just interact between box approaches and the robots in its way.

ROBOTS: Okay, very nice.

INTERVIEWEE 4: This is the first application of interaction with mammals and flying robots.

ROBOTS: Sounds nice. Looking forward to seeing the results. Thanks Maia.

INTERVIEWEE 4: Thank you.

ROBOTS: Hi Alessandro!


ROBOTS: How are you doing?

INTERVIEWEE 5: Fine, and you?

ROBOTS: Why don’t you tell us a little bit about what you are showing the public today.

INTERVIEWEE 5: Okay. So here we have 2 robots, basically a lamprey robot and a salamander robot, which are based on the same modules, and the idea is that they are modular so they have every module that is completely independent and can be interchanged as needed. For example if you have any failure. And the salamander is made of the same modules and with a difference that it has another kind of module for the legs which features the same degrees of freedom as the body for, the movement of the body, but we have the 2 rotating limbs for to have our limbs. And this [?] which is the structure of the salamander and the main purpose of this robot is to use it to study neural-biological systems so we have inside the robot mathematical module of the spinal cord of a salamander and the idea is that as [? the signal retrogrades ?] the rhythmic locomotion is controlled by the spinal cord so we have the brain sending very simple signals to the spinal cord and the spinal cord generating audio saturated signals for the muscles in this support central pattern generators. We have here a mathematical module of the spinal cord of a salamander on some high level inside the robot and the user with the remote control is the brain and sends these very simple signals to the robot and this system inside the robot will generate these lateral signals for all the muscles, all the motors in this case, and make the robot move.

ROBOTS: So you are saying it is modular. Does that mean it is completely distributed or is it somewhat centralized?

INTERVIEWEE 5: This is something that will come. For now the main part is that we generate the main control inside the head, which is this CPG trajectory generator, and this is the centralized part. And then it sends the commands [the set points] to all the modules.

ROBOTS: And in terms of difference, so you say this is the walking salamander and that is the swimming salamander if I can kind of generalize it that way?

INTERVIEWEE 5: This one, it can also swim. It swims like this one. Just this one is more efficient because it does not have the drag of legs. Because salamander swim exactly like lampreys, they just keep the legs at the side of the body and the swim like lampreys.

ROBOTS: So what you are trying to convey here, the concept is actually very complex. You have mathematical model which is trying to model physiological or neurological concepts, which is very complex for the public to understand, per se, so how are you conveying this information to the kids and all the people who are coming to visit here.

INTERVIEWEE 5: Actually, they seem to understand quite easily this concept of having a model of the spinal cord, because you can just explain that like in all vertebras, including humans, the rhythmic locomotion like walking, swimming and so on is generated by the spinal cord and the brain does not do a lot. So basically you do not have to think anything when you walk, you do not have to think what muscle you will move now and so on. That is why it is so complicated to learn how to swim when you do not know. Then the brain is there so you just say that you have the spinal cord model inside the robot and the brain is outside and the concept is rapidly understandable I think. The small children will just see the funny part of the robot moving but that’s all.

ROBOTS: Absolutely. And how do they react when they see the salamander swimming? What is their first emotion?

INTERVIEWEE 5: We do not make this salamander swim today because we have fissures in the covers and it is not completely waterproof, so we make the lamprey swim, and the people are interested and they look at it so on. And all the small children try to touch it like an animal, to caress it and so on.

ROBOTS: Yes, I am sure, very nice. Well thank you very much Alessandro for your time.

INTERVIEWEE 5: Thank you.

ROBOTS: Hi roger!


ROBOTS: How are you doing?

INTERVIEWEE 6: Great, thanks.

ROBOTS: Why don’t you start off by telling us about what you are showing to the public today.

INTERVIEWEE 6: Sure. We brought along 2 systems. One of them is really a research project that is ongoing in our lab and the other one is more an educational system. The research project is about using robotics, actually in our group we use robots as tools and try to get new insights into neural control of movements, so how does the brain controls movements, but also how do people recover after stroke, for example, so neural [rehabilitation], and the research project we brought is an assessment tool so we are using a robot to assess sensory motor function in humans and this is very interesting because, compared to humans, the robot can very precisely quantitatively measure all the interaction parameters so we can track the movement while we are doing a movement and the movement is actually a very functional movement so you have to pick 9 pegs and insert them into 9 holes. So during this movement we assess all the movement parameters so we can reconstruct the movement but we also assess the dynamics so the grip force for example, you have to actually grip and press when you grab the peg and as you transport then insert it and from this we are trying to get out of this quantitative data, we are trying to get new insights into which parameters are most descriptive of certain impairments and which parameters tell us how the patient evolves over time in therapy.

ROBOTS: What you are developing, are these auxiliary tools for medical use or is this really for research to derive conclusions about…?

INTERVIEWEE 6: It is really both. I guess our labs would be unique in this sense is that we work both on the engineering aspects but also we are very much interested in the clinical questions and applications. Actually this device is partially commercial, but we developed the instrument at Handel [Handel Engineering, GmbH] and now we are using the device for those people from our group that go to the clinic, go see the patients, work with the physiotherapists who do these assessments and our group is now also working on the analysis of this data and combining it with other clinical assessment data that we have and with this we are also going after some clinical questions, so we are really interested in the whole spectrum.

ROBOTS: This is actually the project that you are working on for the NCCR robotics, is that?

INTERVIEWEE 6: This is not the project we are working on for the NCCR robotics. The project for the NCCR robotics is actually very similar also very wide scope, we are interested in developing an actuator that could be used in an active knee prosthesis for example. There is one group working on an an active knee prosthesis, [?] Control, and there is also a group here, in Lausanne, working on an assistive orthosis for elderly people to help them in walking. And we are working on the actuator. What we are trying to do is develop an actuator that can behave similar to a real joint, so this is very difficult to do with conventional technology because if you look at the leg for example, I can swing the leg freely and as I step down I have to catch all the weight so this actuator has to change behavior a lot and this is what we are working on, so there we on the engineering side developing this actuator but we are also developing tools to measure how the knee behaves and then trying to use that again to apply it on the actuator and then integrate that with the prosthesis.

ROBOTS: Okay. Are your methods well defined through your hardware solutions or do you work also on software machine learning, etcetera solutions to solve this?

INTERVIEWEE 6: We do a bit also of machine learning. I guess we do a bit of everything. The difficult part is that everything comes together. We are not an advanced machine learning group but we do mechanical design, mechatronic design, system integration, control, and this is control that is very different from industrial robot control, for example, but also the important thing is really to take into account the human factors. So you have to really take into account how the muscle physiology and everything else works.

ROBOTS: So I know that you have worked for a couple of years in Japan. You were active at the ATR in Kyoto and you have also been active in Switzerland for quite a while in research. Do you think that one could maybe say something about the different approaches that these two countries have towards robotics?

INTERVIEWEE 6: Of course one thing that most people are aware of is that Japanese just love robots. People are much more open out in society, and here people are much more worried, and that is something that touches us as well of course because we can develop the most useful assistive system and then if people do not put it on because they feel silly walking around with it outside…

ROBOTS: There is some stigmatization.

INTERVIEWEE 6: Absolutely. And in Japan I guess you have a bigger acceptance so they have also been, from my opinion back then at least, they were investing huge amounts of money into this kind of technology and it was extremely dynamic. At the same time I have to say ATR is a pure research institute so now the University like here from that is also a bit different. But for me it was unique because it was really a research institute that combined robotics and neural-science and this is really how I got into this field.

ROBOTS: Okay. So basically we are talking about society here and how the acceptance of society actually influences the research directions because that is where the money ultimately flows. In that context, what are you doing here now to convey you concepts to the people who are visiting you here are the festival today.

INTERVIEWEE 6: We are trying to make things tangible, that is what we do with haptics, so on the one side they can sit down, try out the system, see what kind of tasks we do and how these robots are used here differently. I guess for a lot of people, if you talk about robots, they still think of industrial robots or humanoid robots. So here they can really sit down and try this out, feel this haptic feedback, and the other thing is we are also showing an educational system that we used already to make our students aware of this multi-disciplinary field of human-machine interaction where they look both at the engineering and the human factor side, so we have this little set up that we show here as well where people can try out these theoretical concepts and feel them.

ROBOTS: And what are their reactions?

INTERVIEWEE 6: I think so far, positive. They love playing around with it, a little more than the equipment can take sometimes, but I think we do quite a few of these events, we try to be present, because it is extremely important that the acceptance has to be there. If it is not there it does not go any further and we will need robots as tools in the clinic in the future because we do not have enough man power and it is all coming to cost.

ROBOTS: Okay. Thank you very much Roger.

INTERVIEWEE 6: Thank you.

ROBOTS: Have a nice day.


ROBOTS: Thank you.

| More