Saturday, December 23, 2006

Paro wins award

BBC news is reporting that the baby-seal therapeutic robot "Paro" has won the won the service prize at the Japanese-government sponsored Robot Awards 2006. Paro is not noew however, I first covered it way back in 2004.

Saturday, August 26, 2006

Nanotech and robots, is Japan the ideal place?

I have recently been reading a bit about future robot technologies, in particular the work of Ray Kurzweil such as "The Age of Spiritual Machines" and his most recent book "The Singularity is Near". There is a lot of interesting stuff in these books about the way future technological innovations and progress will change our lives. Now, you can argue with some of Kurzweil`s predictions and I am sure many of them will be dismissed as fanciful by a lot of readers. However, his prediction about the future development of robotics and AI are interesting. Kurzweil envisages a future where robots will become our physical and intellectual superiors and the only way we will be able to stop from becoming just another evolutionary dead end will not be by trying to compete with the robots (our biological evolution is far to slow to stand a chance of competing with the exponential increase in the information processing capabilities) but rather, we will merge with the robots, initially as human-machine hybrids and, possibly, in the long term, migrating to full machine based life forms. All this will be possible only with advances in nanotechnology. Nanotechnology will be required if robots are to be able to better animals physically (just look at the clunky walking of even the most advanced of the current generation of bipedal robots and compare it to human movement generated by the action of billions of molecular motors to see the benefits of nanoscale technology). Nanotechnology will also be required if robots are ever to reach human levels of intelligence. Now, the question of whether conscious machines can be made and whether artificial intelligence at human or greater than human level is even possible is a huge, contentious debate which cuts across all areas of science and philosophy and is one that I can`t even begin to touch upon here. For those of you who want to get into this argument, one interesting place to start is the "Chinese Room" argument put forward by philosopher John Searle as a way of showing that machines could not posses "strong AI" (a form of artificial intelligence that is conscious and able to reason rather than blindly follow instructions). Reading this argument and counter arguments should provide some basic background to the AI debate. The argument in favour of the eventual emergence of truly intelligent robots goes something like this: Most scientists believe that there is no non-physical component to the brain, i.e. there is no spirit or magical substance that resides in your head, rather everything that makes you you comes from the physical properties of the neurons, chemicals, etc that constitute your brain and that consciousness, thinking etc is simply the "emergent property" of a lot of electrical and chemical signals running through the complex circuitry of your brain. Given that this is the case it should be possible to copy one information processing device (your brain) using a different information processing device (a computer chip). The materials may be different, but so long as they both process information in the same way, the results should be the same. Of course the brain is a pretty complicated organ and so it would require a complicated computer chip to equal it. This is essentially a question of miniaturization. The fundamental component of the computer processor is the transistor. Transistors are continuously shrinking as manufacturers try to fit more and more into their chips and current chip design is already well into the realm of nanotechnology consisting of feature sizes below 100 nanometres. Continuing miniaturization will require new manufacturing technologies but is likely to proceed apace meaning that our computer chips will continue to get faster and faster (for the same price). Finally, in order to interface non biological and biological systems together, we will also need nanotechnology. Trying to interface the miniscule circuitry of a computer chip seamlessly to the tiny features on single neurons is going to require devices on the nanoscale.

So where does Japan fit into all of this? As we all know, Japan is pretty much the world leader in robotics and the physical capabilities its robots show are second to none. It is continuing to make use of its traditional strength in miniaturization to make the robots physically more capable and robust. In addition, Japan has a long history in just the kind of nanotechnology that will be required to produce intelligent robots: The country has long been a leading producer of computer chips and companies such as Toshiba, Panasonic, and NTT continue to carry out cutting-edge research, this is helped by the extremely strong academic research in the appropriate areas of the physical sciences that Japan enjoys. I can see that it is only a matter of time before some bright researchers begin to combine these fields and start producing a newer generation of smarter, faster robots. What does that mean for us bog-standard humans? I`m not sure...

Sunday, June 11, 2006

When all else fails, post a photo of the Hello-Kitty Robot (it can talk with you!)

Sunday, May 14, 2006

Life-like robot MC

There are 3 basic aproaches as to the appearance of a robot: Don`t try to make it look life-like, try to make it somewhat life-like, or try to make it very life-like. The problem with option one is that people tend not to interact with something that is obviously a machine, or at least, their interactions are not very deep. Option two has the so-called "uncanny valley effect" where at first people think that the robot is organic but then, they realise that it is a machine, and the fact that the machine looks partially life-like makes them feel uneasy. The problem with the third option is that it is very difficult to make a robot look very life-like and if it fails, it runs the risk of falling into the uncanny-valley problem mentioned above. If it succeeds it runs into another problem - if a robot looks completely realistic then people will expect realistic behaviour; if your robot looks exactly like a cat, you had better make sure it can do all the things a cat can do, otherwise people will be very disappointed. The robot on the righ above is an attempt to make a robot that looks very life-like. I hesitate to call it a "robot" as it has essentially no autonomy. Nevertheless its face, was very human looking and shows the direction that the external appearance or robots may be taking. Already such robots have been used in a limited role in information booths at expos etc. It won`t be long before we see them at reception desks and info booths at department stores and companies. Apologies for the short length and distance of the video. It was unavoidable due to the crowding at the venue.

Sunday, April 16, 2006

This is a new robot from Tmusk, the RIDC-01. It is about 4 feet tall. It is one of those robots I am not too sure about. It is big, heavy and has a strange combination of functions, the main ones being floor cleaning and the ability to project DVD movies from its head. Of course! An obvious combination. One assumes it is aimed at businesses; very few homes in Japan need such a large floor cleaning `bot. It can also speak and recognise voice commands. How useful those capabilities are in a machine whose main job is to shine the floor and project the occasional movie, I�m not sure. The cost, 10 million yen (50 thousand pounds, 85 thousand US dollars) is a bit steep. I�d rather buy a rhoomba and a nice projector and have enough cash left over for, erm, a sports car. I am not sure if they are on sale yet, if they are, I bet they�re just flying off the shelves.

Saturday, March 25, 2006

Does Japan need a "Manhattan Project" for robots.

Looking at my blog, you can see how advanced the research into producing advanced robots, particularly humanoid robots is in Japan. But still the hope of producing a robot with sufficient mechanical prowess to interact fluently with humans and our world and sufficiently advanced AI to do useful tasks and communicate with us still seems a long way off. For Japan especially, with its falling and ageing population the requirement for such robots in the not-too-distant-future is pressing. Some of the technologies required are already being developed, some seem a long way off. To prepare and integrate all the technologies required will be a mammoth task indeed, but the payoffs potentially huge. I wonder if Japan needs to really treat it like the Manhattan Project or putting a man on the mun, i.e. set an ambitious goal and then really fund it to the hilt... The technologies we will need, as I see it are:

1. Power supplies. Current batteries are too heavy and cannot supply enough power. Their life is too short and their recharge time too long. Fuel cells offer the potential to overcome all of these problems and are already well advanced.

2. Movement. Most robots still move using electric motors, these are too big, too heavy and cannot move quickly enough to give the speedy "bouncy" movements we see in humans and other animals. Artificial muscles perhaps bases on shape-memory alloys offer one solution and are already being researched. It seems quite possible that such technologies may become widespread.

3. Senses. Some robot senses such as vision are already beyond human ability. Others such as the the all-over touch sensitivity of our skin seem more difficult to replicate.

4. Artificial Intelligence. The biggest problem! To work in our wolrd and to interact with us, robots will need an understanding of how our world works and how we humans think. To do this they will need at least some idea of what it means to be human. If you know what it means to be human does this mean you are, at least, partially human? Clearly robots with such abilities are still a long, long way off. This will be the big challenge.

A close up of Nagara III. He looked OK but not especially friendly! I guess this robot will go through quite a lot more prototype stages before it is ready to be sold to the general public.

Sunday, March 12, 2006

This is the "nagara" robot produced with funding from NEDO. It is a protototype that is being designed as a kind of lifestyle partner. A robot that will be able to play with humans, by, for example, kicking a ball.