Of the many articles on robotics recently published in popular journals, one from New Scientist (24 August 2011) struck a chord. The article (only available on-line via subscription) introduces the work of Max Versace and others at Boston University on the Animat program. It follows the modern Rodney Brooks approach, emphasizing the importance of embodiment and environment (see for example ‘Planning is Just a Way of Avoiding What to Do Next’). But there was a delicious twist: it ‘cheats’ at the embodiment. There is no body at all. Everything is a computer simulation, including the environment! New graphics cards allow video game designers to create virtual reality worlds with ease. Versace recognized that this technology can also be used to create the world for a robot. This provides a research environment for robotics that is much more flexible (and faster) than that of the real world with its need for a mechanical interface. (This virtual world for robots could be called ‘Zeroeth Life’, akin to ‘Second Life’ for humans, except it precedes real existence!)
The chord it struck was about my prejudice, perhaps, about much of robotics. As a technologist looking across to a side-discipline, much of robotics looks like playing around with toys. Perhaps it’s ‘robotics envy’: it looks too much fun for this to be serious work! In the words of the song:
“Now that ain’t working, that’s the way you do it
Let me tell you them guys ain’t dumb”
In contrast, the Boston work was giving up on the fun part of having robots wandering around the lab in favour of a dry existence in lines of computer code – so it must be serious work! (I did notice the irony that the virtual reality technology that made the Boston work viable originated from the world of computer games.)
Looking at robotics from an electronics/systems perspective, I feel the mechanics is not that important; you should just live with the mechanics you’ve got and make the best of it. (It may not be much – see what can be done with nothing more than a couple of vibrating motors and an infrared transceiver in the Youtube video of swarming robots from Harvard, below). The real problems of robotics lie at a higher level.
What I don’t get is popular culture’s apparent obsession with humanoid robots, and the sort of work that then gets the attention of the public via the media, for example:
- Kismet, the humanoid head complete with fluffy eyelashes;
- Humanoid lookalikes such as Geminoid-DK, the doppelganger of Danish professor Henrik Scharfe (see below).
This is not serious work. Clothing technology in a human-like skin is gratuitous. (Better to take a Pompidou Centre approach – let’s see all the wires and circuit boards inside.) Instead, as an interested outsider, I have four sources for my passing interest in robotics:
- Cognitive science. We are very constrained in how we tinker around with live beings. Building ‘alternative life’ is a solution to this and is the real acid test : if we can’t replicate the behaviour, we can’t claim to really understand it.
- Practical problem: Personal health. There is a demographic personal health timebomb in the Western world. People are living longer and we can’t count on cheap immigrant labour to look after us in old age in the future. ‘Personal assistant’ robots can potentially solve the problem. See for example the EU FET11 ‘Guardian Angels’ and ‘RoboCom’ projects.
- Practical problem: clean energy harvesting. A solar panel just sitting on a roof is one way to harvest clean energy but is not that efficient in Northern latitudes. Autonomous means of ‘scavenging’ energy where it is available and moving it to where it is needed is another.
- Fun: Personal entertainment and education.
The first three are serious; the last is just for fun – but, even then, it’s a serious sort of fun.
I grew up in the ‘programming’ tradition of the 1980’s, typified by the BBC Micro project: the expectation that we would all need programming skills to some degree to interface with computers. For fun, I want to be building and playing with robots. But just as the home computer revolution 30 years ago produced a generation of programmers, we need to have some (serious) toys for producing a new generation of engineers for them to produce the robots for the serious robot applications (e.g see above). Lego Mindstorms is great but it’s expensive. What is needed is something equivalent to Raspberry Pi which should be trying to relaunch the home computer programming project very soon with a $25 board that plugs into the HDMI port of a TV to turn it into a home PC. So there will be Raspberry Pi in the stocking this Christmas. I then just need to wire it up to a couple of vibrating motors and an infrared transceiver.
Writing this blog post has actually caused me to reconsider my position. I said that ‘personal assistant’ robots can potentially solve problems for an ageing population. But technology typically doesn’t develop the way we expect it to. Visions of the future from 40-odd years ago had robots scurrying around after us, tidying and cleaning. But we do not have general-purpose robots washing-up, nor even a dedicated dishwashing robot clamped to the kitchen sink, with arms to pick up plates, wash them and restack them on the draining board. Technology has instead provided us with the domestic appliance called a dishwasher. The dishwasher is effectively an extension of ourselves, under our control; there is no need for it to be autonomous. Many of the solutions to help the physically infirm in the future may similarly be appliances or tools. What particular need is there for them to be autonomous?
A particular concern with rising healthcare costs issue is the increasing numbers of mentally infirm. The physically infirm need appliances and tools to help them manage – and help them maintain their independence and dignity. However, the mentally infirm will not be able to use the same technology; they will need autonomous agents to make decisions on their behalf, to care for them. But they may also need these caring robots to provide them with social interaction – interaction that a typical human of sound mind would find tiresome or disturbing. And giving such robots a ‘human face’, perhaps even complete with big eyelashes, would clearly help.