In the past two decades, robots have stepped off of the silver screen and into human positions of peril. They lend their helping-if not yet entirely dexterous-hands to dangerous work tasks. Robots have begun their invasion, of the most benevolent kind.
In Japan last spring, technicians working on a laptop watched a feed from inside the tsunami-battered Fukushima Daiichi Nuclear Power Plant. The video, delivered a safe distance away from the radiation threat, came from a pair of PackBots as they rolled through passageways and opened doors to delve deeper into the damaged reactor units and assess the destruction caused by the tsunami. Measurements taken by the robots, which were created by the Bedford, Massachusetts-based iRobot, indicated radiation levels that would have exposed human workers to more than their maximum allowable exposure for an entire year in only a few hours.
Across the Pacific, just outside Oakland, California, a SWAT team tossed what looked like a small, spiked barbell into the open window of a murder suspect's home. The Recon Scout Throwbot, from Edina, Minnesota-based ReconRobotics, Inc., rolled up to the suspect's bedroom door and took a look around the darkened home. Its reconnaissance allowed the team to follow the suspect's movements and safely bring him into custody. The tiny robot also alerted them to the presence of an 8-year-old boy, whom they carefully removed from the situation.
In another extreme, a Yeti rolled over the Greenland ice sheet carrying a load of supplies and equipment. This is a robot, not the mythical abdominal snowman, and its creators hope that it will replace the need for perilous human-driven trips by researchers at the farthest reaches of the island. Its light weight means it is less likely to fall through snow bridges, and on-the-fly intelligence means more efficient-not to mention less stressful-monitoring of fatal ice crevasses along the route. Those who routinely have to make such trips would welcome the peace of mind provided by this robot, which is a combined effort of the Army's Cold Regions Research and Engineering Laboratory, Dartmouth College, and the University of New Hampshire.
These are just a few of the robots that have been introduced and adapted to help keep humans safer. According to the experts in the field, they are only the beginning.
The Sudden Proliferation of Robots
The evolution from the industrial robots on assembly lines of the 1960s to today's creations reflects scientists' ability to give robots awareness of their surroundings, according to Rodney Brooks, an MIT professor and one of the original founders of iRobot. When it comes to stepping into perilous positions, this intelligence makes them perform better and keeps the people around them safer.
The process through which robots become smarter is called machine learning, which involves data sets, sensors and algorithms, and its complexity reflects the brilliance of the computer scientists responsible for it. Progress in artificial intelligence in the past few decades followed a change in approach. Ironically, by not demanding that a robot's computer brain "think" with the same logic as a human, we can ask robots to do the things that humans do.
Advancements in computing have meshed with new engineering techniques to create the robots that today have been assigned to fields that pose the greatest dangers to people working within them, particularly the military. Most other industries-and their risk managers-have limited need for a Predator drone to hunt down insurgents; or a Dragon Runner to hurl out of a moving vehicle so that it can dismantle a roadside bomb; or a SARbot that swims under perilous flotsam and building debris searching for victims after a natural disaster.
But as a society, we absolutely need these kinds of robots. Their creators-San Diego-based General Atomics, QinetiQ North America in McLean, Virginia, and San Diego-based SeaBotix Inc., respectively-and others like them have done humanity a great service. The messages iRobot gets from troops using its bomb disposal units say it all: "Your robot saved my life today."
The robotics field is not limited to these heroics, however. Robots are predicted to solve some of the more mundane but persistent woes of the workforce going forward.
Consider, for example, health care workers. Nurses and orderlies are plagued by back injuries. Each incident may not result in many lost days, but repeat injuries leave many in pain for life and others permanently disabled.
One day, HAL might help. Short for hybrid assistive limb, HAL, created by the Tokyo-based company Cyberdine, is a wearable robot that augments human physical capacity. It could reduce the likelihood, for example, that a nurse strains his or her back while trying to move a patient.
The same technology could be applied to any physically straining job. Honda, also based in Tokyo, has designed a similar technology, the Bodyweight Support Assist. It is currently in the prototype stage, but the designers hope that this tool could help reduce back strain injuries in all types of jobs requiring manual labor.
For now, the risk management utility of robots in the workplace remains largely in going places humans should not. Bill Ross, principal investigator at Carnegie Mellon University's National Robotics Engineering Center (NREC), created the Sensabot to function as a robotic inspector in dangerous settings.
In an oil field in the Northern Caspian Sea, where temperatures can range from 100 degrees to 35 below, it can assess the condition of pipes, valves, pumps, motors and bearings using a variety of sensing capacities. The human operator remains a safe distance away, while the Sensabot works unbothered by toxins, gasses or the bulky protective equipment a person would have to wear.
NREC has also developed the autonomous loading system (ALS), which is essentially a backhoe that works with a remote operator on excavation sites. One safety benefit is a sensor that detects potential hazards as it enters the work area. Moving the human driver out of the cab can also reduce injuries, as these most often occur as operators enter or exit the vehicle, or suffer fatigue-related accidents at the end of a workday.
In a less perilous setting, the NREC Junior (JR) robot spends its days picking up and moving large palettes of plants in agricultural settings. Its improvements to efficiency are complimented by increased safety in the field for an industry heavily dependent on manual labor. For the benefit of the same industry, roboticists at the MIT Computer Science and Artificial Intelligence Laboratory (CSAIL) have been working on what they call "precision agriculture," in which robots deliver individualized care to crops as well as harvest the fruits and vegetables that make for back-breaking human work.
From home and office service robots to more advanced automation in manufacturing, security, transportation, energy and agriculture, roboticists remain confident their creations' future applications will improve the human condition. But why are we still waiting for more robots in everyday industrial settings?
Price remains the main issue. A very expensive machine makes sense only where it can save lives or large sums of money. More robots will thus be introduced to these areas before they trickle down to less treacherous endeavors. And trickle down, we are assured by experts, they will.
Still, businesses may discover ancillary robotics benefits even today, according to Ross. "I think the biggest win for smart new robots is in new applications where the robot enables a new paradigm or a new way of doing business," said Ross. "Robots will allow new methods to be used which were simply not possible before. For example, you can imagine mines which don't need extensive ventilation, cooling, roof-support or safety systems since all the human miners are safely on the surface controlling tough, underground robots. This new approach to mining might provide enormous new efficiencies and cost savings."
But if robotics is to continue the advancements that have brought such possibilities to light-as well as the more distant future of ubiquitous applications across all types of industries-much more work needs to be done.
Robots on the Rise
Robot autonomy inspires public fear, fueled largely by films. Ironically, however, it is the very lack of autonomy that tends to disappoint people who expect an independent electric being when they encounter a robot.
Most roboticists do not aim to recreate C3PO or the other life-like androids Hollywood has made iconic. Although humanoid robots have been developed-Honda's ASIMO, for example-most scientists feel that the field's greatest potential hinges on improving specific capabilities, not necessarily a humanlike appearance or demeanor. There are those, like Brooks, who are attempting to advance the definition of a robot from a machine that performs a task automatically to one that also has an awareness of surroundings. But their aim is generally to create greater intelligence and effectiveness in an automatic-not necessarily autonomous-taskmaster.
Most robots today require some human oversight and control. They also interpret their surroundings and perform tasks on demand-which is much more complex a process than the general public might assume. The fact that they still require human judgment, however, helps in terms of providing inherent safety mechanisms (armed robots cannot choose to fire), financial exposure (you do not want a robot to make a move that might cause it to break) or the need for further performance (i.e., computing) refinement.
These latter two points are considerable research and development challenges for the robotics industry. Roboticists have enough to do without trying to make a robot look human. Trying to design a robot that has simply the dexterity and environmental interpretative ability to open a fridge and grab a pickle jar leaves MIT geniuses awake at night.
Among their many challenges, roboticists need to create more complete and seamless artificial intelligence; dramatically improve dexterity and manipulation; increase reliability; discover enhanced power capacity; establish standards and norms to drive capabilities and industrial efficiencies; develop more suitable materials; and establish a dedicated robotics supply chain. Most of all, they must do all this while creating a cheaper final product. Without lower costs, it will take even longer for robotics to become affordable enough for a broader customer base.
Imagination's Limit
"The world needs robots," said Brooks. This is convenient since his latest company, Heartland Robotics, in Cambridge, Massachusetts, is taking aim at the more affordable variety.
He agrees, as do all roboticists, that robots are not here to take over. They are here to improve upon a human skill, or remove a person from what he calls "the dangerous, the dirty and the just plain boring." The pace of their invasion may not be as fast as some would like, but there has been progress, and with it comes near-limitless options for how robotics can make human lives safer and easier.
"In all cases there will still need to be people to operate, maintain, repair and direct the robots," said Ross. "The number of people [required at a particular job site] might even increase, but safety, flexibility and productivity will increase even more."
The fun of an industry in its "embryonic stage," as McKinney puts it, lies in its limitless possibilities. In three to five years he sees widespread public acceptance and integration, and in 10 years, anything goes. The promise is there; it is simply a matter of thinking of the problems that a robot could solve. And as more problems are presented that need solutions, robots will slowly follow the path of their technological predecessors like the PC or the mobile phone. They will become a ubiquitous, invisible presence that makes each day a little easier.
Therein lies the responsibility for the end-user. Where can you imagine a robot could make your job easier, your workforce safer, your liabilities and accidents lower? Consider what you need and demand it of the industry. You might just get it.
"You start to get into sci-fi quickly," said McKinney, "but that's where it gets exciting."