How cellphone technology drives robotics
An interesting article on Robots.net
highlights one of the factors driving robotics – manufacture of cameras and sensors for ever-smaller cellphones and PDAs.
As an example, the new Sharp LZ0P3721 CCD camera has a 1 megapixel chip (comparable to digital cameras just a couple of years ago) and a 7 frames/second processing rate for about $90 US. Such capabilities cost in excess of $100,000 only a few years ago, according to an article on the second coming of industrial robots.
And it is not only cameras that are getting faster and more portable — gyros, accelerometers and other sensors are also being minaturized, again for the cellphone and PDA markets. Why would a cellphone need a gyro? Current research centers on making cellphones and PDAs more aware of their environment. A cellphone with the proper sensors might know when it was stuffed into someone’s pocket, for example. This knowledge could be used to make the cellphone more responsive to its user and capable of some auto-configuration — in contrast to the traditional PC model in which a device must be manually configured for every task.
If we define the term “robotic” as referring to machines which attempt to sense and react to the real world – as opposed to computers which expend their processing power building artificial “desktop” and “cyberspace” worlds for humans to enter — cellphones and PDAs are on the path to becoming robotic. Because these devices are small and mobile, their sensors must be small, cheap, and run with low power. This translates into a big gain for mobile robots, which can integrate many of these sensors on their somewhat larger bodies.
In this light, a mobile robot is just the next step beyond a smartphone, and wiring up a PDA to control a robot body isn’t as strange as it seems. After all, these devices are evolving with robotic technology integrated from the beginning. Evidence of this may be found in the “hybrid” features of simple wheeled robots like Toshiba’s ApriAlpha. This mobile robot is like a PDA on steroids, in the words of the company:
“Toshiba has developed ApriAlpha to function as a user-friendly interface between the equipment in a home network and the people who use it. This innovative approach positions ApriAlpha as a ‘Robotic Information Home Appliance’ that allows users to effortlessly operate equipment connected to the home network. “
In other words, ApriAlpha combines the function of a home entertainment remote with PDA functions.
In contrast, conventional PCs are not acquiring robotic technology. The main thrust in PC development continues to be how fast their processor runs, and how many “features” their software sports. The basic paradigm for user interaction remains unchanged in 2003 from what it was in 1993 or even 1983. My 3 GHz PC still doesn’t know when I’m standing in front of it, while my 200MHz smartphone does.
As the “robotic” way of programming is explored, it is becoming obvious that brute power isn’t the be-all end-all Recently, Sony announced its upgraded Aibo, The main changes, aside from more streamlined design, are a greater number of sensors. Abio now has sensors on several points of its legs and body, and can use this information to avoid falling over edges, e.g., a table edge or staircase. It also has improved obstacle detection. What’s remarkable is that all of this is being accomplished with a 32mb memory stick – a fraction of the memory put into most computers these days. Like a PDA or cellphone, the Aibo works without running the fastest processors. The new Aibo has a faster processor, but the designers felt that additional sensors were just as or more important.
The real breakthrough will come when sensation is used by the robot in a realistic way. At present, you can pet an Aibo and its responds. However, this is a response devoid of context – the programmers make it respond simply because they want it to do so. In contrast, real dogs respond favorably to petting because it is a form of grooming — which is important because keeping themselves clean is essential to their survival. Once the Aibo is sophisticated enought to “groom” itself (meaning it examines itself and constantly freshens up its appearance by removing scratches and dirt) a pet from a human will be interpreted as a kind of grooming – and a favorable response will be logical in a real-world context. It is strange, but possible to imagine a robotic PDA circa 2010 responding in a similar fashion to being cleaned by its owner.
Looking to the future, it seems that the move to mobile devices currently underway may ultimately be seen as the first wave of the consumer robotics revolution. Future historians may see the cellphone as a kind of simple, first-generation robot, sharply distingushed from personal computers. Since there’s big money in cellphones and PDAs, efforts to minaturize sensors will continue, and in just a few years mobile robots will have a range of cheap, complex, and fine-grained senses. This, more than improvements in speed or programming will make robotics part of everyday life.