Robotic Nation Evidence

One of the teams in DARPA's Grand Challenge race used an innovative technique for object avoidance. This technique shows what is possible as more and more cheap processing power becomes available through Moore's Law. Here what Popular Science magazine had to say about the technique in the June 2004 issue:

Digital Auto Drive of Morgan Hill, California, developed an innovative new robot vision system that, team leaders claim, nearly won them the race.

In theory, the team's strategy is simple: Avoid the struggle to combine data from multiple "active" sensing systems like radar and lidar. Create a real-time 3-D map of the course in front of the vehicle with video cameras alone. Choose a safe path through the upcoming terrain, then point the steering wheel down that path. Now do it while flying through the desert at up to 60 mph. The team, led by engineers David Hall and his brother Bruce, mounted two digital video cameras atop their 2003 Toyota Tundra. A processor transforms the images from the cameras into a 35-billion-pixel terrain map, and redraws that map 60 times a second. For each one of these updates, another processor creates up to 100 possible paths through ifie terrain, then rates each path based on how flat it is, how close it is to the DARPA-defined course, and the size of the obstacles on that path, (The team completely avoids the problem of object recognition by treating every bush like a boulder, and avoiding both.) Repeat the process 60 times a second: 35 billion pixels, 100 paths.

By today's standards, that's a big computational load. Let's assume a simple, efficient algorithm that uses only 10 processor operations per pixel. 35 billion pixels means 350 billion operations. 60 times a second means 21 trillion operations per second.

Obviously that is not how they do it -- even the largest supercomputers today are not much faster, and they are immense. NEC's earth simulator can only support 35 trillion floating point operations per second. It draws megawatts of power to do its job. It certainly would not fit in a truck.

But let's say they do want to use an algorithm that consumes 21 trillion operations per second. Today that takes perhaps 10,000 Pentium 4 chips. In 20 years or less, it will take only 10 chips. And this does not consider the fact that a Pentium chip is general purpose. A single chip will be able to handle the load by tuning it to the vision-processing application. By then, there will also be algorithms that can detect the difference between bushes and rocks 60 times per second. There will be algorithms that can recognize endangered turtles and avoid running over them. And so on.

The thing that is holding back computer vision right now is processing power. As more and more processing power becomes available, innovative algorithms will exploit it. It will not be long before cars can see the road better than people can see it, and our cars and trucks will drive themselves far more safely than people drive them today.

In one of John Dvorak's PC Magazine columns this year, he speaks about humanoid robots in this way:

Creepy News of the Week Dept.: Now Sony and Honda are both making humanlike robots. In the past, Sony has concentrated its technology on creating doglike robots, while Honda has been using its robot to do TV commercials. Now Sony has come up with a humanlike robot that can jump and run, which seems to be incredibly difficult to accomplish. I don't know about you, but these robots are creeping me out. These companies should stop now, before they teach the robots to kill us!

The robots already know how to kill us. Apparently he has not seen this post, this post, this post or this post, nor has he read Manna.

The conventional wisdom says that, as robots take over jobs, the increased productivity should be good for everyone. With robots taking the mundane jobs and increasing productivity, people should make lots more money.

This article from USAToday has a paragraph that paints a different picture:

Since Bush took office, nearly 700,000 manufacturing jobs have disappeared from the region. Many are gone forever, outsourced to places such as China and Mexico where labor costs are lower. Even if the jobs are replaced, the salaries and benefits often aren't. "The jobs that are replacing them are much lower wage," says Jeff Collins, a business professor at the University of Arkansas.

Manufacturing is, right now, the most robotic area of the economy. As the service sector turns robotic, we should expect to see this same downward spiral there as well. Manna talks about this -- most of the "normal" jobs head toward minimum wage instead of getting better as the robots arrive. Instead of the wealth created by robots spreading out to everyone, it concentrates in the wealthy, since the wealthy own the robots.

See also this post.

You would think that "grading essay questions" is something that computers would not be doing anytime soon. Yet they are already doing it now. And students prefer robotic graders to human graders.

Indiana Essays Being Graded by Computers

From the article:

Craig was one of 48,500 Indiana juniors gathering in high schools across the state to take the end-of-year online English essay test. Unlike most essay tests, however, this one is being graded not by a teacher but by a computer.

Craig has already decided he prefers computer grading. "Teachers, you know, they're human, so they have to stumble around telling you what you need to do," he said at a practice session. "A computer can put it in fine print what you did wrong and how to fix it."

Also:

To dispel skepticism over computer scoring, student essays were simultaneously graded by a computer and trained readers during a two-year pilot program. Using artificial intelligence to mimic the grading process of human readers, the computer's automated scoring engine, known as e-rater, generated grades on a six-point scale that were virtually identical to those of the readers.

Teachers -- like doctors and nurses, truck drivers and pilots -- are expensive. So there is huge economic pressure to eliminate them. This new grading technology will advance rapidly, and its combination with computer-based training and robotic teachers will lead to the elimination of several million teaching jobs by 2020 or so.

The fascinating question is this: what will the millions of teachers, truck drivers, nurses and pilots do in this new robotic economy once they lose their jobs?

This article talks about a new system embedded in bar glasses that sends a signal when the glass gets empty.



One thing that this demonstrates is how interesting robotic fauna will become in the near future. Nearly everything will be "intelligent", even if only in a small way.

It is also very easy to imagine this system integrating into the Manna system so that Manna can tell waiters/waitresses which tables to go to when.

This article discusses sensor technologies used in the race:

DARPA's Grand Challenge... Robotics Race or Components Race?

And this site has technical papers on all of the Robots that made it to final qualification:

Team Technical Papers

Robots may protect drinking water from terror attacks

From the article:

A network of underwater robots beaming up a near real-time environmental profile of lakes, rivers and reservoirs could soon be on the prowl helping safeguard the nation's drinking water from sabotage. The robots would replace researchers who painstakingly collect water samples in bottles and take them back to the laboratory for analysis, an expensive, time-consuming and sometimes dangerous practice.

By summer 2005, Syracuse University researchers will have installed a dozen robotic sensors to form the largest underwater monitoring system of its kind in the country and one of the most extensive in the world, said principal investigator Charles Driscoll, a professor of environmental systems engineering at Syracuse.

The title of the article is a little misleading -- the robots would detect problems rather than protect the water. But it is easy to imagine another set of robotic sentries encircling a lake to prevent anyone from entering the watershed.

One thing that this sort of system would help with is industrial pollution. For example, if a sewage pipeline breaks and starts dumping raw sewage into a watershed, or a factory starts dumping chemicals into a waterway, a system like this would detect the problem immediately.

It would appear that we are about to witness the arrival of the robotic hospital. For example, there is more news about doctors using tele-operated robots:

Robot doctor gets thumbs-up from patients



The doctor controls the robot from a console:



From the article:

Eighty percent of the patients taking part in the study thought the robo-doc would increase accessibility to their physician, while 76 percent believed having the robot available would permit physicians to provide more medical information.

"Most patients were very comfortable with this new technology," Kavoussi said.

Kavoussi owns InTouch Health stock and is a paid member of the company's scientific advisory board. The terms of the arrangement are being managed by the Johns Hopkins University in accordance with its conflict-of-interest policies.

InTouch leases the robots for $3,000 a month, which includes computer software, training and technical support. The robots are not for sale because the software is updated about once a month, said Timothy Wright, InTouch's vice president of marketing.

Meanwhile, at the hospital's pharmacy, robots are now delivering medicine to the nurses:

Advanced Technology Allows Robot to Steer Around Obstacles and Call for Elevators

From the article:

A robot named Mr. Gower now navigates the hallways of the R Adams Cowley Shock Trauma Center, riding elevators and opening doors on its own. Its mission is to deliver patient medications to nurses’ stations. The robot, named for the pharmacist in the movie “It’s a Wonderful Life,” has an onboard computer and advanced infrared “light whiskers” allowing it to steer around people and obstacles. The robot even speaks.

So now the robotic pharmacist can pick the medicine out and hand it to the robotic delivery robot. Soon a robot will take medicine directly to the patient.

Combine this with the previously mentioned automated intensive care unit.

Combine it with the previously mentioned robots for physical therapy.

Combine it with the tele-robotic surgeons now emerging in the operating room.

This is all rather remarkable when you consider that, just a year or so ago, hospitals had zero robots. We are about to see an explosion. See Robotic Nation for details.

It was just last week that Microsoft started talking about "dual-core" processors in 2006, and already Intel has made it official:

Intel changes plans, focuses on 'dual-core' chips

If you look at this presentation from Intel, Intel is predicting something like 1.8 billion transistors running at 30 GHZ by 2010. With 1.8 billion transistors, you could put 14 of the latest, greatest Prescott CPUs on a single chip. 30 GHz is about 10 times faster than today's chips. So that is a potential 140x speed improvement over the next six years.

And there are other avenues to high-speed CPUs as well. For example, this paper was featured on Slashdot. It shows how the inexpensive graphics processors found in most PCs today actually have an incredible amount of general-purpose computing power on tap. For certain types of problems, the GPU can be three times faster than the CPU. Once we start creating custom vision processors, custom knowledge processors, etc., things will rapidly accelerate on the robotic hardware front.

The June 2004 issue of Popular Science has a short article on a growing trend in America's intensive care units: remote monitoring of patients. There are video cameras mounted in the patient's room. A remote doctor/nurse watches the patient with the camera and uses a number of electronic monitoring devices attached to the patient. Before turning on the camera, the remote doctor rings a doorbell to let the patient know that he/she is about to be observed.

The system is run by a company called Visicu. Eight hospitals are currently using it, with eight more coming online soon.

Using this system, one doctor/nurse can monitor 50 patients.

The economic pressure to replace doctors and nurses with robots and automation is immense, just as it is for pilots, because doctors and nurses are extremely expensive.

Now that this system has been successfully installed in intensive care units, three things are inevitable:

• Automation will push the number of patients handled by a doctor up. Right now one doctor monitors 50 patients. Soon it will be 75, then 100, and so on.
• The system will expand out from the ICU to normal hospital rooms.
• The doctor will no longer be located in the U.S., but instead will be outsourced to India or China.

This trend has the potential to significantly reduce health care costs, which is good if the savings are passed on to patients rather than being concentrated in CEOs and executives of healthcare companies. However, it also means that the number of doctors and nurses will fall eventually as robots take over healthcare. What will those doctors and nurses end up doing once they are fired?

Humanoids for the Home

Several fascinating quotes in this article, including:

"By 2010, the ministry [of Economy, Trade, and Industry in Japan] hopes, full-fledged, humanoid robots will be common sights in middle-class homes."

And:

Help in negotiating the complex environment of a modern home, enthusiasts argue, will come from a network of tiny radio frequency identification chips. At the National Institute of Advanced Industrial Science and Technology in Tsukuba, scientists are training robots to sort and wash dishes by combining visual data with RFID input. If a robot sees something round and platelike, it scans the object. The plate’s RFID chip reports, in essence, I’m a plate! I get washed and put in the corner cupboard, second shelf!

The Robotic Nation will be here before we know it...

It is fascinating to me that so many articles like this talk about "robots in the home", but never suggest, or even hint, that those same robots will be deployed in the workplace. If a robot can wash dishes, clean, monitor, etc. in the home, then obviously that same robot can work in a restaurant, hotel, prison, hospital, etc. The restructuring of the economy will be remarkable when this transition to robots occurs.

Longhorn to Steal Limelight at WinHEC

This part of the article is amazing:

Microsoft is expected to recommend that the "average" Longhorn PC feature a dual-core CPU running at 4 to 6GHz; a minimum of 2 gigs of RAM; up to a terabyte of storage; a 1 Gbit, built-in, Ethernet-wired port and an 802.11g wireless link; and a graphics processor that runs three times faster than those on the market today.

The first beta of Longhorn will be ready sometime in 2005. Let's assume the code actually ships in late 2006. It is amazing to think that a dual-processor, 6 Ghz machine with 2 GB of RAM and a terabyte of disk space will be "normal" so soon.

A typical PC today (mid-2004) has a single processor running at 3 Ghz, so in 2006 computers will be four times as fast. A typical PC today ships with 256 MB of RAM, so RAM will go up by a factor of 8. A typical machine today ships with a 80 GB disk drive, so disk space will rise by a factor of 12.

That is how quickly hardware is advancing. That is why computers will be reaching parity with the human brain in a remarkably short period of time. Moore's Law continues...

On the other hand, Time magazine points out this week that Longhorn is way off schedule, and is actually hurting the rest of the industry with its delays. Too bad Microsoft is not keeping up with the advances in hardware. There is an excellent post on Microsoft and the damage it is doing available here.

NASA Considers Robot to Save Hubble

The first paragraph of the article is fascinating:

Think Edward Scissorhands, but with bolt drivers and pliers for hands and a giraffelike neck topped with a pair of cameras resembling black cratered eyes instead of the handsome head of Johnny Depp. Could this strange robot take the place of astronauts in fixing the Hubble Space Telescope? NASA is yearning to find out.

One thing that is so interesting about the idea is that most people would consider "repair" to be well beyond the capability of robots. But robots will excel at repair. Robots will repair our automobiles, our homes, our roads -- even themselves. See Robots in 2015 for a discussion.

The Shadow Dextrous Hand :



The hand is powered by compressed air. According to the company: "The Hand is driven from a block of Air Muscles mounted behind the Hand on the 'forearm'", in a manner similar to the human hand (which is also driven by muscles in the forearm). It is quite flexible -- the thumb alone, for example, has five degrees of freedom. "Each joint is sensed individually by Hall effect, with typical accuracy 0.2 degrees."

This paper offers a more detailed description and more photos (including a photo of the air muscles).