Tuesday, August 17, 2004
Robosapien more than a toy - plans hatched in S. Africa
An article on a South African website has some very interesting news about Robosapien, the BEAM robotics biped created by Wowwee toys using technology developed by followers of Mark Tilden. According to the article entitled Humanoids make their debut in KwaZulu-Natal , scientists at the University of KwaZulu-Natal were approached by Tilden to research making Robosapien technology into a non-toy humanoid robot.
Glen Bright, running the Mechatronics and Robotics Research Group in Durban and Auckland plans to produce a 1.65 meter humanoid robot based on Robosapien (similar to Asimo) that can do household tasks and work in healthcare (sound familiar). Bright says that the 1980s were the decade of the PC and the 1990s were the decade of the Internet - and the 2000s will be the decade of robots. I would disagree and say the 2010s will be the 'robotic' decade but otherwise he is on the 'new era' robotic wavelength.
The interesting thing about Robosapien, according to a recent article in Servo Magazine is the methods to make its legs and arms work. Instead of stepper motors attached to each joint controlled by a microprocessor, Robosapien uses a set of sliding plastic parts to deliver output from only a couple of motors. This reduces part number, wear and tear, and generally is more practical for low-cost production than the systems used in the Sony QRIO or Honda Asimo. Motor control in the Robosapien uses the "nervous net" methods of BEAM technology. This greatly reduces the amount of digital computing power used by Robosapien, and makes it more flexible than it might otherwise be.
Interestingly, another consequence of BEAM computing is that Robosapien's power lasts a lot longer than other robots like Sony's Aibo. With 4 batteries it can run many hours longer than systems which rely on digital computers for all their processing and control.
Apparently, Tilden and others have decided that this technology can be adapted quickly to create a "real" humanoid robot. Robosapien is just a stalking-horse for the real thing. Bright says their Robosapien-based system will have a regular computer processor (probably piggyback on the BEAM motion technology), speech recognition, vision recognition, USB and wireless communication. So we will have the typical robotic smarts grafted onto a low-power BEAM-driven body.
Could it work? Despite the limited abilities of Robosapien, it is incredible that it does what it does. No other system remotely offers the ability to walk, move arms, and grab stuff for $90 US - all of which Robosapien can do.
The choice of partners far from the US is also telling. Could it be that the US computer industry is uninterested in developing humanoid robots mixing digital and nervous net technology? If so, it could be that we'll have a robot surprise in a few years - robots that jump coming from India and/or Africa where they were least expected.
An article on a South African website has some very interesting news about Robosapien, the BEAM robotics biped created by Wowwee toys using technology developed by followers of Mark Tilden. According to the article entitled Humanoids make their debut in KwaZulu-Natal , scientists at the University of KwaZulu-Natal were approached by Tilden to research making Robosapien technology into a non-toy humanoid robot.
Glen Bright, running the Mechatronics and Robotics Research Group in Durban and Auckland plans to produce a 1.65 meter humanoid robot based on Robosapien (similar to Asimo) that can do household tasks and work in healthcare (sound familiar). Bright says that the 1980s were the decade of the PC and the 1990s were the decade of the Internet - and the 2000s will be the decade of robots. I would disagree and say the 2010s will be the 'robotic' decade but otherwise he is on the 'new era' robotic wavelength.
The interesting thing about Robosapien, according to a recent article in Servo Magazine is the methods to make its legs and arms work. Instead of stepper motors attached to each joint controlled by a microprocessor, Robosapien uses a set of sliding plastic parts to deliver output from only a couple of motors. This reduces part number, wear and tear, and generally is more practical for low-cost production than the systems used in the Sony QRIO or Honda Asimo. Motor control in the Robosapien uses the "nervous net" methods of BEAM technology. This greatly reduces the amount of digital computing power used by Robosapien, and makes it more flexible than it might otherwise be.
Interestingly, another consequence of BEAM computing is that Robosapien's power lasts a lot longer than other robots like Sony's Aibo. With 4 batteries it can run many hours longer than systems which rely on digital computers for all their processing and control.
Apparently, Tilden and others have decided that this technology can be adapted quickly to create a "real" humanoid robot. Robosapien is just a stalking-horse for the real thing. Bright says their Robosapien-based system will have a regular computer processor (probably piggyback on the BEAM motion technology), speech recognition, vision recognition, USB and wireless communication. So we will have the typical robotic smarts grafted onto a low-power BEAM-driven body.
Could it work? Despite the limited abilities of Robosapien, it is incredible that it does what it does. No other system remotely offers the ability to walk, move arms, and grab stuff for $90 US - all of which Robosapien can do.
The choice of partners far from the US is also telling. Could it be that the US computer industry is uninterested in developing humanoid robots mixing digital and nervous net technology? If so, it could be that we'll have a robot surprise in a few years - robots that jump coming from India and/or Africa where they were least expected.
Monday, August 16, 2004
Robots and the myth of disruptive technology
One of the reasons for the blindness of the traditional PC/Internet industry to robotics is that it fails to fit the criterion of "disruptive technology" - a concept used to sell many a dot-bomb to investors in the 1990s. Simply put, disruptive technologies (conceptualized by Clayton Christensen of Harvard Business School) are supposed to be innovations which allow you to do the same thing, except faster and easier. With the end of the dotbombs, the current poster-child for disruptive technology is Linux, and it is about as close an example as we are going to get. Linux is an operating system that is very similar to Windows, but is (supposedly) less subject to bugs and hacks and very cheap (free). By the rules of disruptive technology, Linux will replace Windows due to these advantages.
The concept of "disruptive technology" has a great appeal to investors. A new company touting its disruptive technology is in effect saying that it can do the same things as its competitors, except faster and cheaper. Its competitive edge comes from improving the product itself, not other competitive advantages like distribution or customer support. So the tech company has a chance of succeeding even if it fails at these traditional tests of business competitiveness.
"Disruptive technology" may exist in software development. Excel replaced Lotus 1-2-3 in part because of easier integration with other popular Microsoft products. The two spreadsheets did the same thing, yet one was faster and easier to use, if not cheaper.
But in an excellent article on ABCnews.com, John Dvorak shows that larger swings in tech do not fit the disruptive tech mold. For example, personal computers were not cheap small minicomputers or fast large calculators. Instead, they were a completely new product concept that expanded independently of these other areas. The spreadsheet, the original "killer app" for PCs, replaced pen and paper with so many new features that it was a completely new concept. The PC was not replacing spreadsheet-like tech with a disruptive faster and cheaper method. Instead, it brought something new to the table.
Over time, minicomputer and calculator sales dropped, but not by being displaced - instead, people moved to using PCs and designed their work around it. In short, the PC was not faster, better, cheaper technology replacing something that existed - it was a completely new bundle of abilities and limitations.
Due to the appeal of disruptive tech to investors (and the chasing of same investors by the PC/Internet world) it has been difficult for many to see the rise of robotics. Even though they share similar components (e.g. wires and integrated circuits), Robots do not allow you do to PC work faster or cheaper - they are something completely new.
This accounts for the myopia most in the PC /Internet word have to robots. When they think of it at all, tech pundits usually imagine robotics as disruptive - voice input replacing keyboard, for example. But voice input is not a faster, cheaper replacement for typing. Even a perfect voice translator will not work like a keyboard. Instead, this technology has a completely different use, enabling computers in situations where keyboards are not used. In short, voice recognition does not replace keyboards.
Similar issues surround the use of robots. When tech pundits imaging robots in the home, they envision them accessing the web (presumably by voice command). It is certainly possible, and if it were faster and cheaper than loading your web browser it would be disruptive. But web surfing is a minor component of any robot which is not robotic - it is PC. It is like expecting television to acquire the feature list of AM radio. TV and radio co-exist, and one is not disruptive to the other.
For this reason we can continue to expect that the vast majority of the tech industry will miss the rise of robotics - right up to the point where some robotics company makes the cover of Time magazine.
One of the reasons for the blindness of the traditional PC/Internet industry to robotics is that it fails to fit the criterion of "disruptive technology" - a concept used to sell many a dot-bomb to investors in the 1990s. Simply put, disruptive technologies (conceptualized by Clayton Christensen of Harvard Business School) are supposed to be innovations which allow you to do the same thing, except faster and easier. With the end of the dotbombs, the current poster-child for disruptive technology is Linux, and it is about as close an example as we are going to get. Linux is an operating system that is very similar to Windows, but is (supposedly) less subject to bugs and hacks and very cheap (free). By the rules of disruptive technology, Linux will replace Windows due to these advantages.
The concept of "disruptive technology" has a great appeal to investors. A new company touting its disruptive technology is in effect saying that it can do the same things as its competitors, except faster and cheaper. Its competitive edge comes from improving the product itself, not other competitive advantages like distribution or customer support. So the tech company has a chance of succeeding even if it fails at these traditional tests of business competitiveness.
"Disruptive technology" may exist in software development. Excel replaced Lotus 1-2-3 in part because of easier integration with other popular Microsoft products. The two spreadsheets did the same thing, yet one was faster and easier to use, if not cheaper.
But in an excellent article on ABCnews.com, John Dvorak shows that larger swings in tech do not fit the disruptive tech mold. For example, personal computers were not cheap small minicomputers or fast large calculators. Instead, they were a completely new product concept that expanded independently of these other areas. The spreadsheet, the original "killer app" for PCs, replaced pen and paper with so many new features that it was a completely new concept. The PC was not replacing spreadsheet-like tech with a disruptive faster and cheaper method. Instead, it brought something new to the table.
Over time, minicomputer and calculator sales dropped, but not by being displaced - instead, people moved to using PCs and designed their work around it. In short, the PC was not faster, better, cheaper technology replacing something that existed - it was a completely new bundle of abilities and limitations.
Due to the appeal of disruptive tech to investors (and the chasing of same investors by the PC/Internet world) it has been difficult for many to see the rise of robotics. Even though they share similar components (e.g. wires and integrated circuits), Robots do not allow you do to PC work faster or cheaper - they are something completely new.
This accounts for the myopia most in the PC /Internet word have to robots. When they think of it at all, tech pundits usually imagine robotics as disruptive - voice input replacing keyboard, for example. But voice input is not a faster, cheaper replacement for typing. Even a perfect voice translator will not work like a keyboard. Instead, this technology has a completely different use, enabling computers in situations where keyboards are not used. In short, voice recognition does not replace keyboards.
Similar issues surround the use of robots. When tech pundits imaging robots in the home, they envision them accessing the web (presumably by voice command). It is certainly possible, and if it were faster and cheaper than loading your web browser it would be disruptive. But web surfing is a minor component of any robot which is not robotic - it is PC. It is like expecting television to acquire the feature list of AM radio. TV and radio co-exist, and one is not disruptive to the other.
For this reason we can continue to expect that the vast majority of the tech industry will miss the rise of robotics - right up to the point where some robotics company makes the cover of Time magazine.
