Rich's Robot Musings
September 13, Darpa Robot Race Begins "At stake is a $1 Million first prize!"
I can't help re-running the paragraph I wrote below about the
2004 Darpa grand challenge. I was much more pessimistic than most
people writing about robotics technology in those days, as I am now.
Guess what? I was right about no team finishing the 2004 challenge
and I will proven right that there will never be a robot with a
brain built on a semiconductor-based computer. Of course, by
"robot" I'm writing about something that might be mistaken
for a human (as Capek intended the word to be understood), not the
myriad of devices and algorithms people call "robots" to
make them sound cool. Seriously, people will call anything a robot
to get attention.
(From March 11, 2004) The publicity for the this race has been brewing for months. The basic challenge is this: a fully-autonomous (no human control) car has to travel a fixed-course 250 miles through the Nevada desert. The quickest time that also averages at least 15 mph wins. As of now, the favorite in the race is Red Zone Robotics (associated with Carnegie Mellon University in Pittsburgh), though my bet is that none of the cars will make it. The Red Zone team has spent roughly $3 million developing their entry; so they are clearly not in it for the prize money. What I find interesting is that the task itself seems so simple. Any human with a $5,000 ATV could do it. This clearly illustrates just how far robotics has to go in terms of approaching human intelligence. I'm sure the Red Zone entry has more sensor and computing power than anyone will be able to cram into a human-scale robot within the next ten years. Marshall Brain predicts that by 2050 we will have $10,000 robots with computing power roughly on-par with humans. He may be right about the raw computing power, but it's contests like this Darpa challenge that make me believe that those robots will have no where near a human's ability to cope with unstructured environments.
September 11, 2018 Why do I care about self-driving cars?
Self-driving cars are not robots, at least according to the original meaning of the word, so why do I care about the state of the art of in self-driving cars. The answer is that robotic cars are a very “clean”
robotics application, at least in terms of the mechanicals and lower-level electronics. The task of making the physical structure of a robot with anywhere near the manipulation and locomotion capabilities of a human is extremely hard. I believe it could probably be done using existing technologies, but that is not a given, and it would take a huge
investment of time and money.
Regardless, self-driving cars don’t have this problem. “Drive by wire” control systems for cars have been common place for years. There’s no need for physical manipulation of steering wheels or brake pedals. Those signals can be sent electronically. Cars are pretty big, at least relative to humans. There’s plenty of room for sensors and computers and the like. There’s also plenty of power, either with giant (again relative to humans) batteries or internal combustion engines.
Self-driving cars really distill the problem to the functioning of the robot’s brain. That’s what makes it so interesting to me. Driving a car is one of the easiest things people do. Extending that
task to anything like human intelligence must be at least ten times harder and probably 100 times harder, or more. If we can’t even make the brains of self-driving cars, why would anyone think we could make the brains for a robot?
September 10, 2018 Why are robotic cars so hard?
On September 25, 2012; Senate Bill 1298 “Vehicles: autonomous vehicles: safety and performance,” which establishes safety and performance standards for cars operated by computers on California roads and highways, was filed with the California Secretary of State. Later that year when the bill was signed into law, Sergey Brin declared, “"You can count on one hand the number of years it will take before ordinary people can experience this.” It’s been more than five years since Mr. Brin made this declaration, and obviously he was optimistic.
Chris Urmson, a leading technologist from the early days of robotic cars, declared his son wouldn’t need a driver’s license because robotic cars would take care of the driving by the time he was 16. Dr. Urmson’s son turns 16 next year. Does anyone think robotic cars will be taking care of the driving by next year?
Here’s a recent quote from Bryan Salesky, another well-known technologist from the early days of robotic cars and the current CEO of Argo AI, “We’re still very much in the early days of making self-driving cars a reality. Those who think fully self-driving vehicles will be ubiquitous on city streets months from now or even in a few years are not well connected to the state of the art or committed to the safe deployment of the technology. For those of us who have been working on the technology for a long time, we’re going to tell you the issue is still really hard, as the systems are as complex as ever.”
What is making replacing humans in cars so hard? Modern cars are already “drive by wire,” so it’s not that we need humans to physically manipulate controls inside the car. Is it a human’s natural sensing superiority over artificial sensors? Absolutely not. Modern robotic cars have cameras, radars, lidars, altimeters, gyroscopes, accelerometers, ultrasonic rage finders, GPS and more. They have far more raw sensing capabilities than humans.
There is no disputing that engineers can build the electronics and mechanicals of robotic cars and fill them with sensors that would put a human to shame. That leaves the “brains” of the car as the missing
link. That’s the part that has me worried. I’ve written many times before that there is no path to human intelligence in a digital computer.
The development path of robotic cars is making me feel correct. The
big tech players have poured tens of billions of dollars into the
problem, yet robotic cars have become beholden to ever-increasing computing power
to enable their technical solution. Unfortunately, the digital revolution is
very long in the tooth. No serious person believes we are going to
keep seeing "Moore's Law" kinds of advances in digital
March 30, 2017 Have the economics changed on tariffs?
Tariffs on imported goods have been in the news lately. Traditional economics says that tariffs are bad because less-expensive imported goods are good for consumers. Goods cost less so people can buy more. It is a simple equation, but I wonder if automation is changing the calculus? Just about all experts in automation agree that increasing unemployment of humans is inevitable and soon (if not already) there won’t be enough jobs for everyone that wants one. I believe global trade and automation have already made jobs across the world a zero-sum game. That is, a job gained in one country necessarily means a job lost in another country. In the United States, there is some basic level of support given to people regardless of whether they are employed. We can argue about the quality of the support, but no one can say that
it is free. The people that are employed must pay for the support. If we take it as a given that jobs have become a zero-sum game, then every imported manufactured good has necessarily caused job loss in this country. Since jobs lost in this way will cost this country money because we must support the unemployed, why doesn’t it make sense to tax the imported items at the border? That way the people considering buying the imported items will
see more of the true cost of the item. If we don’t make the cost of imported items reflect their true cost, aren’t we in fact subsidizing imported manufactured goods through our welfare system? I would love to hear reader’s thoughts on this.
September 17, 2016 Some geeky robotics stuff
I got an email the other day from an undergraduate student asking
about a control interface for a micro-manipulator. They are
currently using an "RC style" joystick with the twist
controlling z-axis motion and up-down/back-forth controlling x and
y. It springs back to neutral when no force is applied. Here is how
Dear Student: I would call what you are working on a Human Machine Interface (HMI). I bet there's a visual component (computer screen, VR goggles, etc.) along with the hand controller part.
The hand controller part is also often called the manual
Frankly, if you only need to control X, Y and Z it's going to be tough to beat a traditional joystick like you already have. My experience is that humans are most precise using the small muscles of their hands and fingers, and that's the scale of a traditional joystick. Some force feedback might be helpful. You could experiment with
that, but I don't recommend a manual controller that is at the scale of whole arm or body
motion to control a micro manipulator.
If you are going to try to design a force feedback manual controller, it needs to be very high bandwidth. The structure needs to be very light, there needs to be no backlash and the actuators need to be backdrivable. The Phantom haptic device http://www.dentsable.com/haptic-phantom-omni.htm is a good example of a design that follows these principles and is at the scale of small movements of the hands and fingers.
You could look at this design and learn from it.
May 5, 2016 Things don't seem to be changing very fast.
I got an email the other day asking if I would contribute to an online story about how different things would be in 2020 "if Moore's law continues to hold true." When I replied that 2020 was only 3.5 years away and Moore doesn't believe his law is holding true anymore, there was no
real response. I didn't participate in this story, but it might be fun to read if it does get written.
Who doesn't like a science fantasy story?
Moore's law specifically speaks to the density of transistors on an integrated circuit, but often is applied to all things
computer. This includes desktop and notebook computers, of course, but also tablets, smart phones, servers, wired and wireless networks, etc. Regardless of whether Moore's law holds true, this technology in 2020 is not going to be much different than it is now in 2016. To me, it is not changing very fast any more. Let's take a look at how it changed between 2012 and 2016. The iPhone 6s is out now, but the iPhone 7 should come out later this year. The iPhone 5 came out in 2012. Sure the iPhone 7 will be better than the iPhone 5, but going from an iPhone 5 to an iPhone 7 isn't going to change anyone's life. The iPhone 9 will likely come out in 2020. Sure it will be better than my iPhone 6s. I hope I can keep my 6s running until 2020. Then I'll get an iPhone 9, but making the move is not going to change my life. Not one bit. I can make the same case with notebook computers, tablet computers and just about every other computing-type technology I can think of. 5G wireless should be rolled-out by 2020. That will be nice, but not that different. The extra bandwidth will just get filled up with ads or other digital debris. Sure, self-driving cars are going to be huge and a great benefit, but that's not going to happen in 2020. I hope to see it in my life time. That might be another 35 years - 2051. As great as they will be, self-driving cars are really just a part of the computer revolution. The computer revolution is very much a
continuation of the semiconductor revolution and all experts I am aware of agree that the semiconductor revolution is very long in the tooth.
To me the fundamental questions are, what will replace
semiconductor technology and when will that happen? Until something
replaces the semiconductors in computing devices, I believe we will
continue to see incremental advances in computing applications, but
they will come more slowly than they have since the middle of
last century when the first transistors were demonstrated. It is
much more fun to pretend otherwise, but I just don't see it.
September 20, 2015 Where are all the robots?
A good friend of mine sent me this link to an article about robotics and wrote, "I’m sure this article is common knowledge to you...” Sadly, this is common knowledge to me because it reminds me of how little progress has been made in the robot part of robotics
over the thirty years since I started in the field as a grad
student. Certainly, the automation and machinery part of robotics has made great strides, but as I’ve written before, there’s
more to a robot than automation and machinery. This article is about one of the best robots in the world and, to quote from the article, “The contraption weighs 407 pounds and is powerful enough to bench-press 150 pounds. It’s hazardous for humans to be around.” That strength to weight ratio is terrible and the contraption is hazardous for humans to be around? That is pretty much exactly where we were thirty years ago. This article starts by listing some of the tasks from the 2015 Darpa Robotics Challenge finals. This competition showcased the best robots in the world. Each of these robots has been under development for at least three years with many millions of dollars in funding per robot. The tasks the robots faced in the challenge were really quite modest. They included walking up a flight of stairs, turning a knob, flipping a switch, and the like. Conceptually these tasks are much simpler than the tasks a fruit fly faces every day. Most of the robots couldn’t complete the tasks, but two of them did. Take a look at the video on YouTube. The robots don’t look much like humans. They certainly don’t show any type of higher-level thinking. As I’ve
written before and unlike a lot of people, I don’t believe robots based on digital computers will ever approach human intelligence. The article above also talks about how much money is currently being invested in robotics by
big players that have a lot of money to invest - Google and Uber being two discussed in the article. I have no doubt that this influx of money will generate even more amazing automation and machinery, but I also have no doubt that
these robots will never approach human intelligence.
July 29, 2014 Robots just keep
coming for our jobs
I've written many times that raising the minimum wage is just
going to accelerate the deployment of robotics and automation to
eliminate these jobs. I think the jobs in the fast food industry are
especially vulnerable. Because I work in robotics and automation I
often think of the social implications of this. What will happen
when the jobs flipping hamburgers are all gone? How will that
teenager learn a work ethic and find a stepping stone to something
better? To jobs flipping hamburgers, I say good riddance. That
teenager would do a lot better spending their time studying their
schoolwork. The same goes for jobs packing boxes, kitting parts,
unloading machines, picking parts off a conveyor belt, etc. Just
preserving busy work for minimum wage is a loser's game and raising
the minimum wage is just going to end the game more quickly.
But how about jobs like taking care of elderly people? Here's an
article in the NY Times that talks about what people in robotics
have been talking about for a long time - that social robots in the
going to be caregivers
for the elderly. I was being interviewed on a radio show several
years ago on this very subject and all I could express is "why
would we want to deploy robots for this?" I had one of the
worst headaches I've ever had after that interview. I can't think of
anything worse for robots to do. There's an emotional aspect to this
kind of work that robots will never be able to fill and robotics
researchers who suggest otherwise are just plain wrong. Another engineer on the show
represented the opinion that there was a shortage of people to do
this kind of work. That's why we needed to deploy robots. The truth
to this statement is that there is a shortage of people willing to
do this kind of work for the money we are currently willing to pay
Of course there is some amount of pay where Americans would be
willing to do this kind of work. The
elderly, however, are often on very tight budgets and don't have the
money to pay for it. People on the left side of the
political spectrum, have written about this issue and suggested that
the answer is government subsidized home care for the elderly. I
very rarely support government subsidies. This includes
corporate welfare and tax breaks for the wealthy. I don't think
subsidies help anyone. At the lower end of the income spectrum,
they disincentivize working, At the higher end of the income spectrum, they just contribute
to income inequality. Over the last ten years the government has
dumped more borrowed money into the economy than in the combined
history of the United States and income inequality has gotten vastly
As for subsidized home care for the elderly, I
guess in this case we don't have to worry about disincentivizing the
elderly. It's not like they would run out and get jobs if it wasn't
for the subsidized caregivers. We already are subsidizing health care
and this is very much a health care issue. On the
conservative side of the political coin is the position that this is the
responsibility of family, church and charitable organizations. This
was a reasonable approach when people didn't live so long, and I
think it is still a reasonable approach for other social issues, but
now care giving for the elderly is just too much for this social network to
Though the use of robots as caregivers for the elderly is an
interesting social and political topic, it is also a hypothetical
one. As I've written many times before,
robots based on digital computers are never going to remotely
approach human intelligence and human intelligence is required to
replace people in caring for the elderly, whether in their homes or in group homes. Sure
there will be devices like glorified smart phones and mechanical
assist devices that will help,
but there won't be robots with human-like intelligence.
July 17, 2014 Super job
opportunity for a Lead Robotics Engineer
General: Infratek is a new technological start-up and university spin-off company focused on providing
innovative, full-spectrum infrastructure management systems. We innovate, integrate, and manufacture autonomous robots, cutting-edge non-destructive evaluation technologies, and software systems that assess and maintain critical infrastructure, to include bridges, roadways, and structures. Having recently been awarded funding, we are hiring engineers to be part of our first commercial launch, bringing research prototypes into industrialized products. As the Lead Robotics Engineer, you will be a key leader of the team.
Duties and Responsibilities:
·Overall responsible for the functional design and technical development of autonomous mobile robotic systems, providing expertise and leadership to the mechanical, electrical, software, and robotics engineers
·Provide system-level guidance on the overall functionality and development of robotic systems, ensuring that validation of customer requirements and verification of system performance parameters
·Lead and facilitate initial planning, concept development, and milestone design reviews
·Assist with hiring of candidates and staff by vetting select resumes and applications and conducting interviews
·Effectively utilize professional or academic networks
Required Skills and Experience:
·Doctorate or Master degree in Computer Science, Computer Engineering, Electrical Engineering, Mechanical Engineering or related discipline
·At least 5 years of experience in the design and development of autonomous systems, robotics systems, and human-robot interfaces
·Industry or research work with the full life cycle of robot development, from concept generation to design and manufacture
·Significant experience related to systems architecture, design, and system integration
·Familiarity with mechanical and electrical design of robotic systems
·Familiarity with core artificial intelligence concepts to include navigation, path planning, sensor fusion, and GUIs
Preferred Skills and Experience:
·Direct experience with autonomous, mobile, “rover-type” robotics
·Direct experience with the full product development lifecycle of robots, from design to delivery
·Experience or familiarity with non-destructive evaluation technology or the integration of robotic systems with infrastructure inspection sensor
Contact Colin O'Neill 856-532-2866 with inquiries.
June 7, 2014 Will robots take
over the world?
Over the years I’ve answered hundreds of email questions from young and old alike. Today I got one that asked if robots fart and another that asked if I thought robots would take over the world. As you might guess, these came from the younger side of the age spectrum. I answer their questions below, but first I have to define what the word “robot” means to me. It’s a bit different than the common definition.
The word Robot comes from a play called “Rossum’s Universal Robots” by the Czech playwright Karel Capek. It premiered in 1921 and became very popular in its day. The play explored what has become a familiar theme in science fiction: humans playing God by creating synthetic humans (robots) in their own image. These robots could be enslaved without guilt because they possessed no souls. Eventually Capek’s robots grew resentful of their enslavement, rose up against humans and destroyed the human race. Though Capek adapted the word robot from the Czech word “robota” meaning forced labor (such as a serf would be forced to perform on the King’s land), the idea of robots is much older.
The notion of human-like creatures created by humans can be traced back to the word Golem that appears in the Old Testament. I’m not going to pretend to understand ancient Hebrew, so I can’t interpret what that word means in the context of the Old Testament, but over the last thousand years the legend of the Golem
evolved. The basic story is that a Rabbi or similar holy person can form a human-like shape out of clay. After speaking or writing the appropriate incantations, the Golem would become animated and serve the Rabbi, typically by performing dangerous tasks or protecting the village against attack. Of course sometimes the Golem would get out of control and turn on its makers. Sound familiar? Frankenstein anyone? Terminator?
These days the word robot is applied to just about any remote controlled or automatic system that does some sort of physical task. The word robot is even applied to non-physical software constructs that roam around on the internet (think search engine robots). I personally thank Joe Engelberger for starting this. In the 1950’s Joe founded Unimation, Inc. which billed itself as the world’s first industrial robotics company. They sold their first “robot” to General Motors in the early 1960’s. This was an automated piece of manufacturing equipment that unloaded hot dies in the factory. To me, that’s no robot. That’s an automated piece of manufacturing equipment, but the name sounded sexy and Unimation was fairly successful.
Finally, we get to my definition of a robot so we can begin answering the questions that started this essay. To me a robot has to be of the same size and shape as a human. That’s what Capek intended when he coined the word. Capek’s robots were so similar to humans that they could be mistaken for humans. In my opinion, if it’s not human-like, it’s not a robot. Interestingly, Capek’s robots were not mechanical creatures, but rather were biological beings that were created in a lab and ran on chemical reactions. Capek doesn’t discuss the robots eating or what they might have eaten, but I assume being biological creations they must have eaten some sort of food. Since they were biological, ate food and ran on chemical reactions; Capek’s robots likely would have suffered from the same residual gaseous emissions that mammals do.
That answers one of these kid's question.
Now we get to the question of whether robots will take over the world. For starters, I don’t believe any robots in the true sense of Capek have been created, yet. Furthermore, I don’t think the robotic machines being designed today will take over the world any time soon. And by soon, I mean not in this century and not in the next. Here’s why I believe that to be true.
People who do say we’ll have robotic machines with brains as powerful as humans generally base their argument on Moore’s law. Moore’s law says that the number of transistors on a computer processor doubles every two years. I have two issues with this line of reasoning. First, there is no guarantee that Moore’s law will continue to hold. It’s not a real law, like the law of gravity. It’s just based on an observation Intel’s co-founder, Gordon Moore, made a long time ago. Second, the interconnections in human brains are much more complex than binary computer connections. The connections in human brains have variable levels, are electro-chemical and involve firing rates. This makes them more like analog signals than digital signals. If we assume the resolution of a connection in a human brain is 10 bits and use a common estimate of 100 trillion connections in a human brain, then we would need a digital computer with ten to the power of ninety transistors! Even if Moore’s law did hold true, it predicts the Sun will burn out before we have a digital computer with the power of a human brain. Basically there is no path from digital computers to the complexity of the human brain.
I don’t know if they will take over the world, but we are probably a lot closer to robots of the kind envisioned by Capek than most people think. I’m talking about cloning. Humans have cloned sheep, cats, dogs, pigs, deer, horses and bulls. It’s only a matter of time before we are able to clone humans. How about the little clone that will be created to provide an organ to replace a child’s damaged liver so the child can live? Is that little clone human? Is it a robot? What if a scientist can clone a person, but render the clone somehow inferior? Maybe it can’t talk, or maybe it can’t comprehend fear or pain? Will that clone have a soul? Will it have any kind of rights or protections? Would it be OK to create an army of clones to fight our wars for us? I don’t know the answer to these questions, but I’m sure we are going to be dealing with them, and likely in my children’s lifetime.
May 17, 2014 Raising
minimum wage will promote robotics
I was browsing around the internet and came upon this article with the title, "Scientists Worry Machines May
Outsmart Man." I realize it's old (from 2009), but the topic still comes up all the time. As I've written before, digital computers are never going to outsmart humans. What about Watson winning
Jeopardy you say? Watson is just a huge computer search engine with speech recognition built in. It's true that Jeopardy questions are often nuanced and contain subtle hints, but when the computer has encyclopedic knowledge, the nuance and subtleties don’t matter. The computer just looks up the answers. As I read the article, I also noticed that the headline was in conflict with the body of the article. Here is what is in the article, "The researchers — leading computer scientists, artificial intelligence researchers and roboticists who met at the Asilomar Conference Grounds on Monterey Bay in California — generally discounted the possibility of highly centralized superintelligences." Here's another line from the article, "The researchers also discussed possible threats to human jobs, like self-driving cars, software-based personal assistants and service robots in the home."
Hello? Do we really need scientists to tell us that automation is going to take human jobs? Seen a lot of bank tellers lately? How about those folks that take your ticket when you leave the parking garage? I'm pretty sure most of those jobs have been lost to automation. How about switch board operators?
Do you even know what those are? At one time there were 350,000 people getting paid to do that job. Now there are none. This is my concern with raising the minimum wage. The higher the minimum wage goes, the faster these
low-level jobs will be lost to automation. I'm all for people earning a wage they can live on, but the Federal
government stepping in and artificially raising the minimum is, as usual, going to have the opposite of the intended effect. It will just accelerate the loss of these minimum wage jobs to automation. There will be one-tenth the number of people working at a McDonalds.
It will be just a few folks to service the machines in the kitchen and one person in front helping
the people who can't figure out how to use the touch screen to order their food.
May 5, 2014 Students explore human-robot interactions
I just read a column about and undergraduate course at Cornell on human-robot interactions. I didn’t
know there were undergraduate courses specifically addressing human-robot interactions, but I’m not surprised. The author writes that the field is so new that there aren’t even textbooks on it yet. I don’t know if that is true or not, but the field is certainly not new. It’s been around since the 1970’s and has focused primarily on safety of personnel working in the same locations as robots. As more and more automation is applied in the consumer market this study will move towards “lay people” that encounter robotics and automation in their daily lives. I’m sure safety will still be an important part of that study, but how to make the interaction “natural” for humans will also be a big part of it.
That leads me to a question I get by email all of the time, “what subjects should I study if I want to work with robots when I grow up?” The obvious answer would be to study math and science, but that’s rarely the answer I give. I think that students should study what they love and are best at. Then they can look for career fields where they can apply themselves. For example, the lead protagonist in Asimov’s “I, Robot” series is Susan Calvin, a robot psychologist. If psychology is your thing and you want to work in robotics, study psychology and figure out how to apply it to robotics and automation. When I was growing up I loved to take things apart, see how they worked, and then try to rebuild them so
they would be better than they were before. That naturally led me to study engineering and I’ve been very happy with that choice, but engineering is not for everyone and it shouldn't be forced on anyone. That's not the way to be happy and successful in life.