What is the future for robotics?

18 September 2012

Interview with

Alan Winfield, University of the West of England

Dominic -   How far are robots from those science fiction depictions

Industrial RobotsAlan -   Well, I'm sorry to say Dominic, some way away. It would be easy to be deceived because at present, we can build some quite remarkably high fidelity android robots, especially if you look at some of the robots from Japan.  But the problem is that their behaviour doesn't match their appearance.  It's what I call 'the brain-body mismatch problem'.  So, those android robots may look terrific, but they're little smarter than a washing machine.

Dominic -   And I guess all these robots are based on computers and we've seen tremendous developments in those computers in the last 25 years.  How has that affected what those robots were able to do?

Alan -   Well, surprisingly little and that may shock you.  You're absolutely right that of course, we have extraordinary levels of processing power, and we can put incredible computing power into the robots.  The problem is of course that the computing power is just a raw material.  It's how you use it that's important and solving the fundamental problems of artificial intelligence - in particular, general AI - the ability to solve general problems, for instance, still eludes us.  So, that's a problem of designing the AI so, until we solve that problem, no amount of computing power is going to help.

Dominic -   And I guess a factory is a fairly well-controlled environment where you know what kinds of objects the robot is going to encounter whereas a home is a much more complicated place.

Alan -   Well, that's exactly right.  I mean, as you said in your introduction Dominic, the transition that's going on right now is from robots in factories and warehouses, and undersea exploration, and so on, to personal robots.  To robots that will actually be workplace assistants or companions - therapeutic robots in the home.  But it's fair to say that we're still at the beginning of that transition.

Dominic -   I guess there are also safety issues here because in the factory, you can fence the robot off, but in a home, you really don't want it pulling someone's arm off.

Alan -   Well, that's exactly right.  I mean, one of the major problems that's being addressed in the field of human robotic interaction right now is robot safety.  It remains a very challenging and unsolved problem. And as you say, not just the physical, as it were, strength of the robot is a problem but also, the fact that the environment is completely unpredictable.  Building a robot that can both deal with those uncertainties and be completely safe, utterly safe around children for instance is a big challenge in robotics.

Dominic -   You mentioned safety around children.  What sort of applications are you envisaging and do these safety concerns affect your choice of applications?

Alan -   Well, I mean, I mentioned children simply because children might be present in the home.  I mean, some people have spoken of robots that might for instance help, look after children, childcare robots, but personally, I think that's going too far.  I'm one of those roboticists who think that some areas of human activity should not be replaced by robots, and I think childcare is one of them.

Dominic -   And I guess also looking after the elderly might be a similar problem.

Alan -   Well, indeed and care for the elderly is something that's particularly pressing in the Far East, and Japan and South Korea already are very advanced in healthcare-assistant robots for the elderly.  But I have to say that it's a problem that we're facing too in Europe.  Again, very difficult technical challenges, but also, the ethical questions are those that we need to be concerned about.  In other words, would it be ethical to - for instance - replace a human carer with a robot?

Dominic -   What do you think about using robots in education?  I guess there's two questions there because it's whether they can be used to teach and this question that whether children could learn about science and technology by playing with robots.

Alan -   Well, absolutely.  I think that the second of those points you just made is absolutely wonderful.  There's no doubt that robots are an extraordinary, if you like, tool for education.  I mean, in UWE for instance, we've been using robots for probably 15 years to help teach programming to undergraduate students.  But as for robots actually being teachers, again, I can see roles for such robots as  an assistant or offline teacher, but certainly not replacing human teachers. For young children, it would be great if their 'robo-teddy' for instance might be able to read them stories, and there's a kind of educational function in that.

Dominic -   So, looking at head, what do you expect to see in the future from robots in the next 10, 20 years?

Alan -   Well, that's a hard question.  Certainly, there's a whole range of research that's going on right now in human robot interactions.  So for instance here in the lab - Bristol Robotic Lab, here in Bristol - so we have work for instance on gestural communication so that robots can understand human body language.  We have research in object recognition and handling, so that if you ask a robot for instance to pass you the cup of coffee, it can actually recognise the cup and then of course safely, pass it to you.  Other areas I think that are slightly more far out, but I think are equally important are new materials. So robots need to be lighter, they need to be compliant and soft.  Then if all else fails and the robot falls on you, it won't do you any damage because it's too light, for instance.  And then the area of robot self-awareness, for instance. I think that these future companion robots will have to have a level of self-awareness in order to function properly and do their job.  I think it's worth saying as well, Dominic, that putting together an enormous range of, at the moment, separate pieces of robotics research is a grand challenge in robotics.

Dominic -   Thank you very much.  That was Professor Alan Winfield from the University of the West of England in Bristol.

Chris -   Alan, you're going along to see Colin's boats?

Alan -   Well, I'd like to very much.  I know the group at Aberystwyth well. Terrific, good luck with the race, guys.

Chris -   I've had a question for you from Catherine Hiscox and she says, from Hemel Hempstead, "Will robots be able to think for themselves?"  What do you think?

Alan -   Yeah, it's a good question.  The problem is, of course, we don't actually know what thinking is.  But I would say in principle, yes, they should be able to think for themselves, but it's going to be a long way into the future I'm afraid.

Chris -   Thanks very much, Alan.

Add a comment

CAPTCHA
This question is for testing whether or not you are a human visitor and to prevent automated spam submissions.