Credit: Photolibrary
FOR RODNEY BROOKS, the path to becoming one of the world's foremost roboticists began in the backyard shed of his childhood home. It was Glenelg, Adelaide, in the 1960s, the space race was in full swing and Brooks and Scott Johnston, his mate from around the corner, spent endless hours in that shed trying to blow things up.
Brooks' father, a former telephone technician who worked at Woomera, Australia's long-range missile testing facility in the desert 500 km north of Adelaide, brought home leftover rocket equipment to aid their childhood fantasies.
Their aim was to make solid fuel rockets, but mostly they made smoke bombs. Using their own recipe, they produced vast quantities of hydrogen and learnt that a garbage bag full would rattle the windows of the neighbours' house when exploded.
They spent all their pocket money buying junk from a local equipment recycler, built a hovercraft that didn't hover and an oscilloscope - a shock from which once caused Brooks to pass out, requiring his brother to revive him.
"It's not that we ever succeeded in finishing any of these lofty projects," says Johnston, now a high school physics teacher. "More that we had to problem-solve and improvise relentlessly." Skills that have been vital for Brooks as he's evolved from backyard boy wonder to 'bad boy' of the robotics world.
According to his colleagues, no one has been more successful at seeing into the future of robotics, at defining new ways to build, think and bring robots into everyday life than Brooks.
He's described variously as a "robot revolutionary" a "gigantic presence in the robotics field" and a "renaissance man" with a "hugely powerful mind". He's no tin man without a heart either; friends describe him as a "classic connector", an "emotional and romantic man".
He has been twice married, is passionate about his children (Brooks is father to four - three biological children and a stepdaughter, all now in their twenties) and his dogs (he's the owner of two Yorkshire terriers whom he says have taught him more about robot behaviour than being a parent ever has). He also has a soft spot for all the robots he's had a hand in creating and reels off their names like members of his extended family. He has a heightened empathy for machines - he winced watching the film Robocop when the robot went belly up, because so many of his own robots have done the same.
Former students at the Massachusetts Institute of Technology (MIT), in Cambridge, which Brooks recently left after 24 years of teaching, describe him as a mentor "not just as a scientist but also a citizen". One says he thinks of Brooks with the same affection as his favourite uncle.
A great raconteur, a hush always fell over the crowd when the Panasonic Professor of Robotics entered the room. His classes were popular because they were fun, students say. He's also a provocateur; he once ran a series of lectures called "God and Computers" which caused great controversy on the MIT campus - many of his former students remember them as the most interesting and thought provoking lectures they attended at the institute. He's the author of a host of books, including a science fiction novel that's yet to see the light of day.
Brooks thinks of his work as a grand adventure, which is perhaps why some call him a workaholic. He's also a savvy entrepreneur and a multimillionaire.
He brought the world its first robotic vacuum cleaner, the Roomba, which at last count had sold four million; he also delivered the PackBot, the radio-controlled robot now used by the U.S. military as a forward scout in Iraq and Afghanistan, estimated to have saved thousands of lives.
In a rare combination, he's equally adept at cutting computer code as he is at wowing an audience with his flair for public speaking. An irreverent man of many dimensions, he loves scuba diving, marathon running (he's run the Bos ton marathon three times - at the back of the pack) and beautiful architecture (he worked with architect Frank Gehry on the building that housed his MIT laboratory).
But in order to understand his own natural intelligence, which is much more layered and complex than the Artificial Intelligence he strives to manufacture, we need to go back to that corrugated iron backyard shed where Brooks' love affair with building machines began.
The extraordinary life of roboticist Rodney Allen Brooks had very ordinary Australian beginnings. He was born in 1954, the second in a family of four children who grew up in a cream brick South Australian suburban home.
He swam at the beach, played cricket at the Chappell Brothers' Saturday clinic ("I was lousy at it," Brooks confesses) and Aussie rules football (which by all accounts he was more unlucky at than lousy, breaking his arm in a game the day before he left Australia to study in the USA).
By the time he was a student at Glenelg Primary, the local public school both his parents had attended, his exceptional mathematical abilities had already been recognised.
"I was a little weird from a young age," says Brooks, whose exceptional intelligence perplexed his parents, Harry, the telephone technician, and Jean, a hairdresser. By the time he got to his final year at high school his maths teacher told him not to bother coming to class - unless he taught it. Even Brooks wasn't sure where his talent could take him.
"I knew I liked blowing up stuff and that I would build computers, I just didn't quite know how to get there," he says.
A defining moment in his career trajectory occurred when the 14-year-old Brooks went to a 1969 Adelaide screening of 2001: A Space Odyssey. He says he was enthralled by the soft-spoken computer villain HAL 9000 and wondered whether we could one day achieve that level of Artificial Intelligence.
Brooks decided he wanted to be part of that effort, so he went on to study pure mathematics at Flinders University at the undergraduate and masters level. His professors recognised his talent, but it was the 12 hours Brooks spent every Sunday on the ancient university computer that were the most instructive - he taught himself how to decode hardcore computer language and wrote his own.
Realising he wasn't going to get anywhere on the world stage if he stayed in Adelaide, Brooks applied to three universities in the U.S. to do his doctorate. He got accepted to two. "I went to the library and got an atlas and looked up where they were. Stanford was in California and it seemed the closest so I went there."
Brooks arrived at Stanford in 1977, in the early days of Silicon Valley, which for a young computer scientist was like journeying to Mecca. It was here the young scientist, with his long-haired hippie look, began to develop his reputation as the 'bad boy' of robotics, challenging the status quo of that burgeoning industry.
Artificial Intelligence, or AI, a phrase coined in 1956, was based on the idea that human intelligence could be simulated by a machine. Back then, AI was the domain of an elite group of university trained young men. "These young guys in 1956, known for their intelligence and wanting to recreate it artificially, looked at what they were good at and decided it was chess playing that distinguished them as being intelligent."
This bias blinded them, says Brooks. He took their idea and turned it on its head (a habit he's renowned for to this day), arguing that intelligence was built layer upon layer from the bottom up, not the top down. "They thought chess playing was the epitome of AI. But I thought elephants were pretty darn intelligent and they can't play chess."
Instead, Brooks felt intelligence was all about the stuff we take for granted, the things a two-year-old can do like distinguish between a pair of glasses and a bottle.
"In the early 1980s, Rod had an epiphany," explains Daniela Rus, director of the MIT Computer Science and Artificial Intelligence Laboratory's Centre for Robotics, and a former colleague of Brooks. "He realised that robots were large, slow, expensive and unreliable.
It was not uncommon for a robot to spend five minutes to decide how to move five centimetres. Rod took on this problem and turned it upside down. He observed that since robots could not construct useful maps and reason about these maps in a timely way, robots should be built avoiding these requirements."
In fact, Brooks decided robots should be more like insects, equipped with simple control mechanisms tuned to their environments. So he set about studying the way ants move.
In the summer of 1988, he was able to put his theories to the test and spent what he describes as "the best 12 weeks of his life" working with the ant-like Genghis robot. After some 300 failed attempts, Genghis moved - it was Brooks' eureka moment in computer science, and lead to his appointment as a tenured professor at MIT.
Colin Angle was an undergraduate student working with Brooks that summer. He says that prior to Genghis, roboticists were dealing with huge computers and a very long series of challenging problems that needed to be solved to get a robot to move. "Rod came up with a way to solve these problems with algorithms that were about 100 times less confusing than anyone else had come up with. It was a breakthrough for us all."
His idea became known as 'subsumption architecture', and it met major resistance from others in the field. Colleagues said Brooks was nuts, that he was throwing away his career.
Instead, his work proved to be a paradigm shift in AI. He argued against 'symbolic processing' approaches - which consider intelligence a matter of processing symbols according to a fixed set of rules - in favour of 'biologically-inspired' machines that can adapt to dynamic environments.
Instead of reinventing mobility in a computer, he looked to biology for an inspiration to questions of robot locomotion. Since then, subsumption architecture has inspired and taught several generations of roboticists how to build machines, and has led to the creation of robots such as Allen, Tom and Jerry, Herbert, Toto and others It also formed the basis for some influential studies, most notably Brooks' "Elephants Don't Play Chess".
If a robot had a human-like body, Brooks observed, people would interact with it in a human-like way, and if a robot had a non-gender based name, his students talked to it differently.
So with his Humanoid Robotics Group at MIT, Brooks began to create robots with faces, and with names like Kismet and Cog. Brooks confesses Genghis remains his favourite robot, but HAL, he says, has stood the test of time (though he thinks it would have been better if HAL had had a body).
Brooks is on the committee of the robot hall of fame at Carnegie Mellon University, in Pittsburgh, which allowed him to correspond with one of his heroes, Arthur C. Clarke, the now deceased author of 2001: A Space Odyssey.
BROOKS BECAME A cult figure with the 1997 release of Fast, Cheap & Out of Control, a documentary by Errol Morris about four very different men and their obsessions.
It also lead to Brooks meeting his second wife, Janet Sonenberg, a fellow MIT professor in theatre arts, who wrote to Brooks telling him how much she disliked the film. (It was not an uninformed critique - her brother is Oscar-winning documentary producer David Sonenberg.) The movie also took his theories about robots to a wider audience, something Brooks had been working at quietly since 1990.
Brooks had always set up companies on the side during his years in university research. And he had good contacts with the U.S. military, which had funded the bulk of computer science research since his days back at Stanford in the 1970s. So after that breakthrough summer working with Genghis, Brooks and then-graduate students Angle and Helen Greiner started a company with the idea of launching the first private mission to the Moon and selling the movie rights.
The idea failed. But it did illustrate Brooks' enthusiasm for robots: the sky was no limit. Instead of putting robots on the Moon, however, the group began to set their sights on bringing robots into homes. Angle, who is now CEO and chairman of the company iRobot, says they always knew that a robot that did the vacuuming would sell. So using a variant of subsumption architecture as their business model (building from the bottom up) they worked with a toy company to learn how to manufacture in China, and with a cleaning company to learn how to clean.
Finally, in 2002 - after 12 years of development - they had a product, the Roomba, which is now the most common robot in homes around the world. Other iRobot devices range from swimming pool cleaners to small robots that can climb stairs and detect hazardous chemicals.
Very visible has been the PackBot, first used in the ruins of September 11 in New York to search for survivors in the rubble, and one of many machines iRobot has developed for military purposes. The PackBot EOD (explosive ordnance disposal) - controlled by radio to handle potential explosives - has been popular with soldiers in Iraq. The company went public in 2005.
Angle says Brooks taught him important lessons - not just as a robotics engineer - but as an entrepreneur. "Rod used to hang rejection letters on his door at MIT from submissions of papers he was trying to get published. Some of the rejection letters were incredibly nasty. He was called a hobbyist with no place in academia. The idea was to show us he wasn't afraid of rejection, he actually embraced it. He would view them almost as badges of pride, medals of honour. He was espousing this idea even before it became a strategy in AI - that it is important to challenge the currently held ideas. And he proved a good role model for us as entrepreneurs, because as an entrepreneur you usually fail more than you succeed."
Brooks recently stepped down from his role as Chief Technical Officer at iRobot. Angle says Brooks didn't always take directions from his former student well.
"Rod could do in a day what it would take one of my programmers a whole month to complete. It was just impossible to predict which day he would do it. Rod is a free spirit. He's a great business partner but he's not easily focused. He enjoys this role of being a little bit of the outsider - the guy who looks under different rocks from everybody else. This strategy doesn't always work of course, but when it does you usually discover something that's important."
In September 2008, Brooks took a leave of absence from MIT and launched a new company, Heartland Robotics. While the details remain under wraps, Brooks says he aims to use robots to do for manufacturing what the personal computer did for office work; make it more efficient. He believes the time is ripe for robotics to take off as an industry.
"From virtually no mobile robots deployed anywhere in the world six years ago, we now have thousands on active duty in the U.S. military and millions cleaning the floors of homes. This is the lead-up to a classic 'hockey-stick' growth curve," Brooks said recently.
"Just as computers we interact with personally transformed our lives over the last 25 years, so, too, will robots transform our lives over the coming 25."
Now 54, his once long hair is short, grey and slicked back - there's a slight resemblance to another South Australian, former Premier Don Dunstan. Ann Whittaker, his executive assistant since 2005, who is working with him as co-founder of Heartland Robotics, says despite his more distinguished appearance, the bad boy title is still appropriate.
"It's been a while since he's flexed his radical ideas. He's been in an administrative role now for many years and is just now preparing to unleash his mature and ever boundary-pushing 'bad boy'. I think Bad Boy 2.0 will be the champagne version. Or better yet, a well-aged red wine version," she says.
BROOKS IS BUOYANT about the future of robotics - especially in places such as Japan where they are developing robots as elder care companions. This is mainly out of necessity, says Brooks, because of the country's ageing population and its lack of an immigrant workforce. To raise the standard of living, he says, people need to exploit a labour force - so why not robots?
This of course raises moral questions. The word robot comes from the Czech robota, meaning forced labour or drudgery. Brooks points out that the degree of AI granted to a robot should be commensurate with the task you want it to perform. Obviously you would want your robot companion to have more intelligence than your fridge, he says. "The question is how much feeling should you give them if you believe robots can have feelings?"
The robotic companion idea seems to have greater acceptance in Japan he says, where there is a belief in animism, that objects can have a soul. "People find it icky to think of robots as companions in the U.S., Europe or Australia. But I believe we are all machines. Biomolecular machines."
One of his dreams is to develop a robot that people feel bad about switching off. As he wrote in his book Flesh and Machines, "We had better be careful just what we build, because we might end up liking them, and then we will be morally responsible for their well-being. Sort of like children."
Do robots have personalities? Will we ever create human-level AI? Brooks answers yes on both counts. The question of the soul is a little trickier, and one he continues to discuss with his childhood friend Johnston. "There have been some things which have never changed about Rod … I think we connect much the same ways we did as kids, still laugh at the same stupid things and we still find things like consciousness intriguing," says Johnston.
Over the years, Brooks and Johnston have been thrilled to watch how well their children got along. And a penchant for blowing things up seems to be an inherited trait in both families. Once, when his second son Andrew was a teenager, Brooks had to pull him aside - he'd been trying to make dangerous devices like Molotov cocktails. Instead, Brooks showed him how to make and blow up hydrogen like Johnston and he had done as kids. "Pops, you're not supposed to be showing me how to blow things up, you're a dad," Andrew said.
He's a dad, a roboticist, an atheist, an evolutionist … and a 'creationist' of sorts. He believes in creating things: robots, rockets or your own wild imaginings in the backyard shed.
Helen Pitt is a freelance journalist based in San Francisco.
