Until now, visitors of the FMA could use a magic wand like the one in the right picture to send quadrotors racing through the 10x10x10m space. As shown in the video, the addition of a Kinect now allows a far more natural and intuitive interaction.
Autonomous cars behaving themselves during the DARPA Urban Challenge
It's one thing to ramble on (like we do) about how autonomous cars are way safer than human driven cars, but it's another thing to prove it. Like, mathematically. A research group at Carnegie Mellon has created a distributed control system for autonomous highway driving and then verified that it's safe. In other words, the software itself provably cannot cause an accident.
To do this, the CMU group started with a simulation of just two cars (equipped with sensors and short range inter-vehicle communications) in a lane, and then proved that their software kept those cars from having an accident 100 percent of the time. With this as a base, they slowly expanded the simulation, adding more and more layers like multiple cars and lane changes until they had an entire complex autonomous control system, each module of which is definitely safe.
So far, the system is only able to deal with entering, exiting, speed changes, and lane changes on straight line highways, so it's going to be of limited use unless you live in Kansas. It's also dependent on sensor technology that is only just starting to be introduced into vehicles, and I imagine that the "provably" bit starts to break down when dealing with unexpected situations, like a moose jumping off of an overpass onto the hood of your car. But it's a start, and a fundamental technique that can be built upon.
This type of thing also seems like it may have the potential to streamline the introduction of autonomous cars from an insurance and legal standpoint, since it offers some degree of protection for manufacturers: If an accident occurs and the software provably cannot be at fault, that leaves either a sensor hardware failure, or (more likely) a human simply pushed the wrong button.
UPenn's GRASP Lab has done some absolutely spectacular things with quadrotors. They've taught them to fly in formation, build structures, and even fling themselves through narrow windows. But as with every learning process, there are going to be some cases where things just don't go quite right, and when you're dealing with fast moving autonomous robots with four spinning rotors apiece, sometimes tests can be, to quote UPenn, "spectacularly unsuccessful":
Spectacular is right. I'm not sure what exactly caused the attempted backstab at 0:20 or the mass suicide at 0:28, but all those violent quadrotor deaths were pretty funny to see. You know, for science.
Oh, and props to the GRASP Lab quadrotor team for sharing their failures along with their successes.
Well, this has to be one of the weirdest things I've seen in a while. It's a robotic armpit, created by Kevin Grennan, a London-based designer. As with human armpits, it sweats. The idea is to use odors to improve human-robot interaction.
For example, you'd mount this thing onto a bomb-disposal robot (like a Talon, in the illustration below), and whenever you sent it in to disarm an explosive, the robot would start sweating, filling the air with the same sort of chemicals found in human sweat when we're afraid.
This may not the first robotic device to emulate the human underarm, but it's definitely the most repulsive and freakish looking. Grennan did this on purpose, explaining that he's interested in how "more complex and private parts of the human body would be translated onto [a] robot."
Here's how he describes the project:
Each robot that I have augmented with a 'sweat gland' emits a particular chemical that has a specific effect on humans and the chemical has been chosen to further enable the robot's primary function.
In the case of the bomb disposal robot the 'sweat gland' releases the smell of human fear. It has been proven that humans can identify this specific smell and it tends to enhance cognitive performance. I propose that this robot would enable surrounding humans to work more effectively and to differentiate dangerous situations from false alarms.
He came up with a couple more sweating robot concepts, too:
The one on the left is a picker robot; the one on the right, a surgical robot. Grennan explains:
In the case of the picker robot. It releases a chemical called androstadienone, which is found in male sweat. This has be shown in research to effect mood in females under certain circumstances. I have speculated that this robot when used on a production line could enhance the performance of female employees in it's vicinity.
The third robot is a surgery robot. It releases a mist of oxytocin, a chemical found in the human brain. This chemical when inhaled nasally has been shown to cause people to become more trusting. I speculated that a patient could meet this robot before surgery and the chemical mist would cause the patient to trust in its abilities to a greater degree.
What's the point of all this? Grennan's goal, I suspect, is not so much to create a functional product as to provoke thought. As he puts it: "I hope that the dark thought of robots taking subconscious control of humans [via their emitted odors] will cause viewers to reflect on how we really want to interact with these machines in the future."
One of the things that I like about robots is that they take some of those features that are intrinsically biological (i.e. fear, and for that matter, stinkyness) and do away with them completely. That said, it's interesting to see the ways that robots are getting to be more like humans just as humans are getting to bemore like robots. If you plotted those two trends on a graph along with an axis for time, at some point you'd find an intersection. As to just what exactly that's going to entail, well, your guess is as good as mine.
We've been wondering why no car company has ever tied lane-assist (which automatically keeps your car in its lane) to adaptive cruise control (which automatically keeps your car a safe distance from cars ahead) to create a fully autonomous highway cruise mode for cars. We're talking about sensor systems that have existed in midrange to high end cars for quite a while (like, years), suggesting that the hold-up hasn't been so much technological as social and legal.
Finally, Volkswagen (who has beena consistentinnovator in the autonomous car field) seems to have decided to do the obvious and let lane-assist and adaptive cruise team up to handle your car on the highway without you needing to do anything at all:
The Temporary Auto Pilot (TAP) bundles semi-automatic functions, i.e. functions monitored by the driver, with other driver assistance systems. It combines such automatic systems as ACC adaptive cruise control, the Lane Assist lane-keeping system and Side Assist lane-changing monitoring into one comprehensive function.
In the semi-automatic driving mode – referred to as Pilot Mode, for short – TAP maintains a safe distance to the vehicle ahead, drives at a speed selected by the driver, reduces this speed as necessary before a bend, and maintains the vehicle's central position with respect to lane markers. The system also observes overtaking rules and speed limits. Stop and start driving manoeuvres in traffic jams are also automated. With TAP, it is possible to drive at speeds of up to 130 kilometres per hour on motorways or similar roads. Drivers must still continually focus their attention on the road, so that they can intervene in safety-critical situations at any time.
Sounds awesome, right? It is, except for that (by now obligatory) last sentence that negates one of the primary benefits of an autonomous system: You have to be paying attention to everything you would normally be paying attention to if you were actually driving, meaning that while the car might technically be doing the work for you, the benefit that you gain is severely limited. You can't take a nap, you can't do work, you can't watch a movie. In fact, you can't really relax. And Volkswagen is quite explicit about the situation, as they state in their press release: "the driver always retains driving responsibility and is always in control."
That line was clearly written by a lawyer. In fact, what Volkswagen is doing here is basically just trying to make sure that they're legally covered, and I can't really blame them for it. If they were to say, "hey guess what, your car is driving now, have a beer and go play Angry Birds on your iPad for a while" they'd most likely be sued into oblivion if something went wrong while the car was in control. At the same time, there will come a point at which the benefits of not having that driving responsibility will outweigh the risks, especially as technology gets more reliable (and traffic gets worse).
It's important to note that giving drivers time to space out is definitely the only benefit to an autonomous cruise control system. Volkswagen rightly focuses on the potential for making cars exponentially safer by using all of these fancy sensors in a more active capacity: Even if you're technically in control, your car is still about a thousand times quicker than you when it comes to reaction times, and especially in heavy traffic situations, TAP has the potential to significantly reduce most minor (and likely even major) accidents. You remember Shelley, right? This is what it's all about: Teaching your car to be a better driver than you, and then giving it the chance to keep you out of accidents.
It's surprisingly easy for humans to endow robots with personalities. We've seen it happen most poignantly with EOD robots, but it's a common occurrence for people with domestic robots as well. However, these robots were never designed to have personalities. They're designed to do a job, and they're designed to be able to interact with people to the extent that it facilitates their ability to do that job, but service robots are really not programmed to be your pet, your best friend, or a member of your family.
Whether it's in their programming or not is, to some extent, beside the point, since it happens anyway. And when it happens, it dramatically changes the way that people interact with what on a primary level is intended to be little more than a tool. Realizing this, a team from Delft University of Technology and Philips Research in the Netherlands decided to take a look at how people actually want their robot vacuums to behave, and what kinds of personalities they'd like them to display.
To do this, the researchers used what's called the Five Factor Model to describe a set of thirty hypothetical personality traits to a group of study participants. The aforementioned Five Factors are broadly described as openness, conscientiousness, extraversion, agreeableness, and neuroticism, and each one of these categories can be subdivided into more specific characteristics like "calm," "talkative," "likes routines," "bold," and "systematic." Each participant was asked to rate how important these characteristics were, and the results were incorporated into a sort of hypothetical "desirable" personality for a robot vacuum.
The next step was to take those desirable personality characteristics and turn them into robot behaviors, and this is where it starts to get a little, uh, strange:
"The translation from personality to behavior was inspired by a role play in which a group of actors was asked to act like a robot vacuum cleaner with these desired characteristics. Attributes, such as macaroni, were available to support acting out some of the situations (e.g. ‘cleaning a dirty spot’). An introductory exercise was meant to familiarize the actors with the personality. Then, the actors were asked to act out situations—as if they were the robot vacuum cleaner—making use of motion and sound (expression through light was taken into consideration only after this exercise). In general, the actors either crawled about or walked around at a slow pace to imitate a vacuum cleaner. Often, a typical vacuuming sound was simulated by them."
I'm sure it's impossible to imagine how hilarious that must have been. And I absolutely asked the researchers for the video but they won't give it to me, I imagine because it would ruin the careers of any of those (I would have to assume aspiring) actors. Sad.
Anyway, they took all of those performances and used them to create their own "prototype" video of a hypothetical vacuum cleaner exhibiting some of the personality traits displayed by the actors. The word "prototype" is in quotes because this is just a little remote control vacuumy-looking thing with the sound dubbed in, but watch the video and see what you think about the personality of this little guy:
A panel of fifteen people were asked what they thought about the prototype, and they were able to successfully describe those personality characteristics that were originally instilled into the prototype, suggesting that it's definitely possible to give household robots personalities, even if they don't have any expressive features beyond movement, sound, and a few blinking lights.
It's not just that it's possible to do create a robot with a personality, but what's relevant is it actually makes a difference to the end user. This is a more important point than you might think; by way of example, consider the difference between the iRobot Roomba and the Neato XV-11. Which one of these vacuums cleans better is certainly up for debate (and we've debated it), but as we've pointed out in the past, iRobot has a perception problem with their pseudo-random method of cleaning versus the Neato's straight line technique. The XV-11 just seems smarter to people, whether or not it actually does a better job, and that makes a difference when people are deciding what vacuum they want to buy.
There's lots of nifty graphs and charts and stuff in the actual paper, which is entitled "Robot Vacuum Cleaner Personality and Behavior," by Bram Hendriks, Bernt Meerbeek, Stella Boess, Steffen Pauws, and Marieke Sonneveld from Delft University of Technology and Philips Research. You can read it in its entirety at the link below.
Just a year ago robotics research startup Willow Garage sent its first class of 11 PR2 robots into research labs around the world. The company's intent at that time was to jumpstart the personal robotics industry. While it indicated that eventually it might want to participate in that industry and sell robots for profit itself, that time seemed hazy, somewhere in a distant future. In the meantime, the company would continue developing open source software and giving a lot of robots away for free or at cost.
But suddenly, the folks at Willow Garage are talking a lot more like business people instead of research scientists. I wouldn't say they've put aside their research hats entirely, but they're certainly thinking about commercialization a lot more than they were last summer.
"A year ago we weren't sure if we would sell PR2s or not," said CEO Steve Cousins, standing next to a PR2 in the company’s tree-filled courtyard earlier this week [photo, above]. Cousins was hosting a dinner for journalists; the robots, by the way, did not cook, though they are capable of preparing a few dishes, like a sausage breakfast and cookies. The PR2 was the company's first robot, a $400,000 platform for robotics research. Since then, the company sold 25 PR2s; it also introduced a $1300 robot kit, the Turtlebot, that uses an iRobot base and a Kinect 3D sensor [shown below with Brian Gerkey, director of open source development].
And now, says Cousins, Willow Garage is putting some serious thought into figuring out just where the market for personal robots will be in the next few years.
The company doesn't expect to sell PR2s in great quantity; Cousins likened the PR2 to the Alto, developed at Xerox PARC in the 1970s. The Alto never became a commercial product, but seeded what became the personal computer industry. And, after an initial batch of orders is filled, Willow Garage will get out of the Turtlebot manufacturing business, letting other companies offer products that follow Willow Garage's design.
Instead, Willow Garage's first real commercial robots are likely to be descendants of the company's Texai personal presence robot. Texai was not intended to be a commercial robot, but rather was instigated by a telecommuting researcher for his personal use. The company initially considered Texai a distraction from its main goal, which was completing the PR2. "It caused a real tension," Cousins said, "but we didn't kill it, and that turned out to be a good decision."
Just a month or two ago, Willow Garage spun out a company called Suitable Technologies, intended to bring the Texai, or some version of it, to market. Willow Garage founder Scott Hassan is heading up the new company, which is busily hiring engineers, and, according to its website, will have its first products on the market in early 2012. And Suitable may not be the only Willow Garage spinout, Cousins hinted; instead, the company will likely seed other new businesses.
Suitable Technologies is coming a bit late to the telepresence party, but the company thinks Texai's large screen and other features will give it an edge. A number of other companies have started selling personal robots. Anybots, also in Silicon Valley, offers its QB system. Vgo Communications, near Boston, has the Vgo robot. And iRobot has been showing off a prototype called AVA. Willow Garage sees this competition not as worrisome, but rather as a sign that its efforts to accelerate the development of a personal robotics industry are working.
Cousins said that when Willow Garage started four years ago, he predicted that the point at which the growth rate of personal robotics would dramatically swing upwards, becoming exponential instead of incremental, was 10 to 15 years in the future. Now, he says, "we've been at this for four years, and we are now five to six years out. So I think we've made a difference."
In a visit today to Carnegie Mellon University's National Robotics Engineering Center, Obama launched the Advanced Manufacturing Partnership, a $500 million program to bring together industry, universities, and government to invest in emerging technologies that can improve manufacturing and create new businesses and jobs.
Robots are a big part of this effort. The administration's new National Robotics Initiative seeks to advance "next generation robotics." The focus is on robots that can work closely with humans—helping factory workers, healthcare providers, soldiers, surgeons, and astronauts to carry out tasks.
"You might not know this, but one of my responsibilities as commander-in-chief is to keep an eye on robots," Obama quipped at the beginning of his speech. "And I'm pleased to report that the robots you manufacture here seem peaceful. At least for now."
The National Robotics Initiative involves the National Science Foundation, the National Institutes of Health, NASA, and the Department of Agriculture, which combined will make available up to $70 million per year to fund new robotics projects.
Obama tours the National Robotics Engineering Center at Carnegie Mellon. Photo: White House
Obama said he visited a local company, RedZone Robotics, which makes a robot that explores water and sewer pipes [photo, below]. "It is fascinating stuff," he said. "It can go through any sewer system. It’s operated remotely by the municipal worker. It’s got a camera attached so it can film everything that it’s seeing." Obama said the robot could potentially save cities millions in infrastructure costs and create jobs for workers to operate the robots and pore through the data collected.
I asked Dr. Henrik Christensen, a robotics professor at Georgia Tech who helped to establish the new initiative, what technical challenges they plan to attack. He explains that traditional robots are good at tasks that require precision and repetition, but to work alongside human workers the robots need to be smarter and safer.
"If we want to build a robot coworker, we need to have safer systems, new materials, better sensors and actuators," he says.
One example is a robot that can observe a human worker performing a task—say, assembling parts of a car—and replicate that task, with the human supervising and assisting the robot if necessary. Another example is a robot that can help workers pack goods in a pallet or truck in an optimal way, to improve logistics.
This theme recognizes the emerging mechanical, electrical and software technologies that will make the next generation of robotic systems able to safely co-exist in close proximity to humans in the pursuit of mundane, dangerous, precise or expensive tasks. Co-robots will need to establish a symbiotic relationship with their human partners, each leveraging their relative strengths in the planning and performance of a task. This means, among other things, that for broad diffusion, access, and use (and hence, to achieve societal impacts), co-robots must be relatively cheap, easy to use, and available anywhere. As the US population ages and becomes more culturally and linguistically diverse, these co-robots may serve to increase the efficiency, productivity and safety of individuals in all activities and phases of life, and their ubiquitous deployment has the potential to measurably improve the state of national health, education and learning, personal and public safety, security, the character and composition of a heterogeneous workforce, and the economy, more generally.
"Our understanding of the applications of robotics technology has expanded," says Dr. Matt Mason, director of Carnegie Mellon's Robotics Institute. "Robots are not just for dull, dirty, and dangerous jobs. Their greatest value is working with people." He says that in addition to reinventing manufacturing, robots will play ever more important roles in agriculture, medicine, rehabilitation, and elder care.
Obama with a robot built by RedZone Robotics for inspecting water and sewer pipelines. Ron Wolf and Sub Vallapuzha of RedZone show the President how the robot can be remotely controlled. Photo: Tim Kaulen/Carnegie Mellon University
And why start a new initiative when agencies are already funding cutting-edge robotics projects? Dr. Christensen of Georgia Tech says robotics research has been too fragmented; what is needed now is coordination.
"The most important thing is that we keep track of the results and make sure that innovations get transitioned to companies."
He says that claims that robots take jobs away from people don't take into account the fact that robots—and other technologies, for that matter—not only help keep companies competitive but also allow them to grow and hire more people.
The White House Office of Science and Technology Policy, which was involved in spearheading the new initiative, listed several reasons to make robotics a priority:
Robotics can address a broad range of national needs such as advanced manufacturing, logistics, services, transportation, homeland security, defense, medicine, healthcare, space exploration, environmental monitoring, and agriculture;
Robotics technology is reaching a “tipping point” and is poised for explosive growth because of improvements in core technologies such as microprocessors, sensors, and algorithms;
Robotics can play an important role in science, technology, engineering and mathematics (STEM) education because it encourages hands-on learning and the integration of science, engineering, and creative thinking; and
Members of the research community such as the Computing Community Consortium and program managers in key sciences have developed a shared vision and an ambitious technical agenda for developing next-generation robotic systems that can safely work with humans and augment human capabilities.
But of course, the real reason behind the initiative: Obama loves robots.
Obama left his signature on the Sensabot, a robotic vehicle designed to monitor dangerous environments. Photo: Byron Spice/Carnegie Mellon University
Watch the speech:
Here are more reactions issued by leading roboticists:
“Investing in robotics is more than just money for research and development, it is a vehicle to transform American lives and revitalize the American economy. Indeed, we are at a critical juncture where we are seeing robotics transition from the laboratory to generate new businesses, create jobs and confront the important challenges facing our nation. The nation’s robotics community is collectively poised to advance the technology and at the same time accelerate the transition of these technologies from the lab into the market.” — Helen Greiner, president and CEO, CyPhy Works; president, Robotics Technology Consortium; co-founder, iRobot
“A well coordinated National Robotics Initiative will provide the leadership and investment necessary to accelerate the advancement of next generation robotics technologies and knock down the regulatory and technological barriers necessary to lead to wide spread adoption and repatriation of US manufacturing jobs.” —John Dulchinos, president and CEO, Adept Technology
“If we want to end the talk of a 'jobless recovery,' we should increase our national investments in robotics and create millions of high-paying jobs in the process. How? By developing improved robotics technology that can be applied to reviving our manufacturing industries, protecting the environment, reducing our dependence on foreign oil and helping provide quality care for our growing elderly population.” — Jeff Burnstein, president, Robotics Industry Association
"The United States has pioneered research in Robotics and Automation; we are now poised to develop the next wave of robots that will dramatically enhance the abilities of human workers in factories, labs, hospitals, and homes." — Ken Goldberg, professor of engineering and information, U.C. Berkeley
"Robotics is a game-changer! President Obama's National Robotics Initiative will lead to new innovative technologies that will invigorate America's manufacturing economy by creating new opportunities and new jobs, improve our quality of life by revolutionizing health care and medicine, and make our nation safer with the development of robots for defense, security and emergency response." — Vijay Kumar, professor, University of Pennsylvania
"The use of robotics in medicine and health care has seen substantial growth in the last decade. Telerobotic systems are routinely used to perform surgery, resulting in shorter recovery times and more reliable outcomes for patients. Robotic rehabilitation systems are delivering physical and occupational therapy, enabling treatment that is continuously adaptable to a patient's needs. In the United States, robots are stimulating the development of new treatments for a wide variety of diseases and disorders, which will improve both the standard and accessibility of care." —Allison Okamura, professor of mechanical engineering, Stanford University; Maja Mataric, professor of computer science, neuroscience, and pediatrics, University of Southern California
"RE2 Inc. is representative of small robotics engineering businesses across the United States that are contributing to the economic growth of their regions. These small businesses form the backbone of the new economy by leveraging federal and private investments to create new high-tech jobs and ensure that the United States remains a leader of global innovation." — Jorgen Pedersen, president and CEO, RE2
Infrared view of NASA's robotic lunar lander prototype
NASA has been working on a prototype robotic lander designed to explore the moon, asteroids, comets, and other airless bodies that don't demand complex and scary aeroshells and parachutes for safe landings. The lander, which seems to have no sexier name than "robotic lander testbed," is about the size of a golf cart and is powered by hydrogen peroxide catalyst engines.
These engines run 90 percent hydrogen peroxide, which is just like the stuff that you can buy at the store, except way more concentrated, making this lander one of the more eco-friendly spaceships out there. This overview video from a few months ago takes you through the prototype:
Just this past Monday, the training wheels officially came off, and the robotic lander took itself to an autonomous two meter hover for 27 seconds. The video below shows the entire flight (along with some touching audio of the jubilant crew at the end) from several angles along with some sweet infrared footage:
Part of the deal with the robotic lander testbed is to try to develop a versatile platform that can be used to field missions that are relatively inexpensive and efficient, and from the sound of things, they're off to a good start. In the future, we may see robotic landers like this landing on the moon to check for volatiles (like water), listening for moonquakes, or even hitching rides on near-Earth asteroids.
On Tuesday morning, a Northrop Grumman MQ-8 Fire Scout was shot down by Moammar Gadhafi's forces in Libya, becoming NATO's first combat casualty in the conflict. The U.S. Navy has been testing Fire Scouts for five years or so, and the robots have progressed from shipboard autonomous landings all the way to accidental drug busts, but this is the first we've heard of them actually involved in a major military operation.
Presumably, the Fire Scouts are being used solely as surveillance platforms, although they've also been successfully tested as weapons platforms, as you can see in the second half of this vid:
So far, the Navy hasn't said much about what exactly the Fire Scout was doing when it was shot down beyond the obligatorily vague "performing intelligence, surveillance and reconnaissance over Libya to monitor pro-Gadhafi forces threatening the civilian population." It's remarkable how not big of a deal this incident is relative to what the response would have been had (say) a manned Apache gunship been shot down instead, as Libyan state TV originally claimed.