|
|
Rich's Robot Musings
April 18, 2012, Cool robot
video
This video is really cool. The electrical, mechanical and software engineering challenges that had to be solved to make this happen were enormous. Ever time I see a video like this, though, it makes me think just how far away we really are from having robots that even approach human's mental and physical capabilities. To me, even a fruit fly is more sophisticated than these robots.
Fruit flies fly around on their own, find food, mate and reproduce.
That's way more than even the most sophisticated robot can do.
I've written before about IBM's Watson computer that competed on Jeopardy. Most folks commenting on the show fell all over themselves marveling at the accomplishment, but frankly I wasn't impressed. 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. IBM spent a reported $1 billion developing a huge encyclopedia that could
do speech recognition? Really? This is nowhere near human intelligence and
aptly demonstrates how far away we are from developing artificial intelligence. As I’ve written before, I don’t believe there is a path from digital computers to human intelligence and here’s why. Watson had 90 processors with about 1.2 billion electrical connections per processor. That gives a total of 108 billion connections.
The human brain has about 100 trillion connections. While it is possible, and even likely, that computers with 100 trillion connections will be developed during the next decade; that doesn’t mean that these computers will have the capabilities of the human brain. This is
because connections in digital computers are binary, ones and zeros. The connections in the human brain are more like analog signals. They are electro-chemical and involve firing
rates. If we assume the resolution of a connection in a human brain is 10 bits, then we would need a computer with ten to the power of ninety
transistors to even come close! The sun is going to burn out before that happens.
March 18, 2012 Here I go
again...
A military engineer (Master Explosive Ordnance Disposal (EOD) Technician)
wrote me a question about robotics and it made me think, some about
robotics and some about how great our men and women in the military
are.
Hey Soldier:
Robotics is an extremely broad field. By definition it encompasses the work of electrical engineers, mechanical engineers, software engineers and computer scientists. If you go just a bit further it gets into social systems, sociologists and the humanities. I always find it interesting that the lead character in Asimov’s books on robots is
a psychologist named Susan Calvin.
You might like to read the play “Rossum’s Universal Robots.” That’s where the word “robot” first entered our lexicon. Before that there is a history of robotics (at least thinking about robotics) that goes back beyond the Old Testament. Do a little searching on the word “golem” and you will see what I mean. I could go on about humans creating humans in their own image, but that probably isn’t why you wrote.
You asked about your first job after “retiring” from the military. You like electro-mechanical systems, so pick something in electrical, mechanical or software engineering. Of those, software engineers have the easiest time finding jobs, then electrical and then mechanical. I think the biggest advances in robotics will come in the area of computer science. As I’ve written before, I don’t believe there is a path to human (or even dog) level intelligence in digital computers. Work on what you like though, that’s what where you will be the most successful.
Take care,
Rich
Feb 11, 2012 This looks amazingly
fun. I may have to apply...
The Discovery Channel is looking for America's most creative and daring
techies, machinists, inventors and engineers to design, build, and BLAST their way to a Grand Prize on their new competition TV show TOP
ENGI
January 14, 2012, I can't
believe it is 2012!
Here are some job opportunities. Shoot Brent an email if you are
interested VanKuiken.Brent@hru-tech.com
Major Duties and Responsibilities (Robot Programmer):
· Create new robot path programs using the robot teach pendant.
· Touch up and teach others how to touch up robot programs using the robot teach pendant.
· Download offline robot programs.
· Assist in robot dressing.
· Assist in robot interface debug.
Major Duties and Responsibilities (Simulator):
· Operate Deneb IGRIP simulation tools.
· Build geometric models.
· Perform robot reach studies.
· Perform robot collision studies.
· Create off-line robot programs
Skills and Abilities:
· Demonstrated ability to solve engineering problems.
· Ability to teach others.
· Proficient at supervising diverse people.
· Proficient at robot programming or robot simulation.
Qualifications
· Bachelors degree in an engineering field or equivalent years of direct experience with a automotive manufacture or supplier.
· Previous GM certification preferred but not required
December 18, 2011, A
student looking for an internship
I received an email from a student at the University of Cape Town. He writes well, seems genuine and is looking for an internship. His CV is strong. I'm sure he'll be happy to send it to you. Here's what he wrote.
"My fields of interest include, amongst other things, automation, CIM systems, complex digital systems, Artificial and Computational Intelligence, automotive control technologies and general data planning and co-ordination of complex systems. My degree programme is in Mechatronics – it's a hybrid branch of engineering, a mix between electrical and mechanical engineering. It differs from the Electro-mechanical degree by focusing more on the electronic control, design and automation of systems."
Send an email (frxchr011@myuct.ac.za)
November 20, 2011, A nice
site with information about robotics
Here is a permanent, updatable repository and invaluable resource for all robotics
enthusiasts. It includes information from labs at MIT, Oxford, ANU, Cambridge, and many German, Canadian and Asian
Universities. You could spend some very-interesting months going through the site.
It's called expo21xx. I suppose that's because it will be beyond the
year 2100 before we have robots anywhere near the intelligence of
humans.
I looked at the page from Osaka University. It talks about the "uncanny valley." My son first told me about this concept. We are all fine with automation, until it gets too human-like. I've asked the question many times: why build an artificial
human? There are plenty of humans around.
October 29, 2011, I love
questions from kids
Hello Dr. Hooper!
Obviously I am planning on being a robotic engineer. I know you get many emails on this subject so I hope I won't trouble you by asking some questions. Because I am in 6th grade some of my peers think that it is a far-fetched idea, however they don't doubt I could do it since me along with my best rival are in gifted and talented. I at first wanted to be a scientist, than I read about robotic engineering, I just would like to know how much of a difference is there between a robotic engineer and a scientist? And also what subjects must I major in besides math and science? Can I study to be both?
Thanks for your time and I hope I hear back from you soon!
Hello young student whose name has been redacted:
I would say the greatest difference between engineers and scientists is that engineers are working on problems that need to be solved in the next year or two and scientists are working on problems to be solved in the next decade or two (or maybe century or two).
As far as choices of classes, study what you enjoy and let the rest take care of itself.
Good luck,
Rich
October 25, 2011, IBM's
Watson
I've been meaning to write about the Watson computer on Jeopardy ever since I watched the shows, but am just now getting to it. Most folks commenting on the show fell all over themselves marveling at the accomplishment, but frankly I wasn't impressed. 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.
Really? IBM spent a reported $1 billion developing a huge encyclopedia that could understand English? This is nowhere near human intelligence and really demonstrates how far away we are from developing artificial intelligence. As I’ve written before, I don’t believe there is a path from digital computers to human intelligence and here’s why. Watson had 90 processors with about 1.2 billion electrical connections per processor. That gives a total of 108 billion connections and the human brain has about 100 trillion connections. While it is possible, and even likely, that computers with 100 trillion connections will be developed during the next decade; that doesn’t mean that these computers will have the capabilities of the human brain. This is because connections in digital computers are binary. The connections in the human brain are more like analog signals. They are electro-chemical and involve firing rate. If we assume the resolution of a connection in a human brain is 10 bits, then we would need a computer with ten to the power of ninety transistors! The sun is going to burn out before that happens.
July 2, 2011, Mobile Robot
Job Opportunity
Gibbs & Cox, Inc. is one of the nation’s leading independent naval architecture, marine engineering and design firms. We are seeking a well-qualified individual for a position as a Senior Robotics Engineer: The Senior Robotics Engineer should have a PhD degree in Robotics or Mechanical PhD with specialization in robotics. Degree should be from an accredited college or university and a target of ten years experience in the development of mobile robotics systems. Demonstrated knowledge and application of perception and artificial intelligence is desired.
Contact - lepstein@gibbscox.com
April 26, 2011, Another very interesting job opportunity
MaryJane Dow | Technical Sourcer | Intuitive Surgical, Inc.
1266 Kifer Road, Bldg. 103 Sunnyvale, CA 94086
Mechanical Engineer – 437391 - TR
Job Location: Sunnyvale, CA
Primary Function:
This position will play an important technical role during the rapid production of a novel new surgical robot instrumentation. The senior engineer will be expected to work as a key member of a team that quickly conceptualizes new mechanical designs, develops them, and brings them to market. The engineer will play a lead role in the design and development of new surgical instruments and accessories, as well as interfacing with manufacturing to facilitate manufacturing process development. The successful candidate must excel in a high-energy team environment and be capable of making sound decisions when faced with the time pressures and incomplete information typical of new product development.
DESCRIPTION
Roles and Responsibilities:
This position has responsibility and authority for:
· Gathering, interpreting and organizing clinical customer input for new designs
· Generation, design and development of new device concepts focused on those needs
· Concept documentation, preparation and review of patent submissions, and support for regulatory submissions
· Detailed prototype design, build, test, and iteration
· Design for cost and manufacturability
· Generation of formal design documentation, review, and control (e.g. design reviews, ECOs, mechanical drawings, etc.)
· Project planning and management
· Vendor selection and management
· Manufacturing transition and support
ADDITIONAL REQUIREMENTS
Competency Requirements
Competency is based on: education, training, skills and experience. In order to adequately perform the responsibilities of this position the individual must:
Skill/Job Requirements:
· Masters degree in ME, or PhD
· An industry-wide reputation for design excellence and rapid product development
· Passion for creating robust and reliable products
· Demonstrated history of products reaching the marketplace
· At least eight (8) years experience with mechanism design, including medical device design experience
· Minimum two (2) years experience bringing medical products to market
· At least three (3) years experience leading projects
· Experience leading small engineering teams and mentoring other engineers
· Knowledge of ISO requirements and GMP guidelines. Experience with FDA regulations and medical device design control a strong plus
· Proficient in solid modeling, SolidWorks preferred
· Familiarity with sterilization processes and designing for sterilization
· Working knowledge of small part assembly and manufacture
March 26, 2011,
This looks like a super-interesting job opportunity.
Dear Dr. Hooper,
I am a recruiter for the Wyss Institute@Harvard in Boston, MA and came across your blog. I was wondering if you would have any interest in posting a job opportunity I have open at the Wyss to your blog and if you knew of anyone who might be interested. Here are the details to the position:
Lead Senior Staff Engineer - Bioinspired Robotics
Duties & Responsibilities: The mission of the Wyss Institute for Biologically Inspired Engineering at Harvard University is to transform human healthcare and the environment by emulating the way nature builds. Developed as an alliance between Harvard and other premier academic and clinical partner institutions, Institute faculty and staff collaborate in high-risk, fundamental research and science-driven technology development. A major focus of the Institute is to translate the technologies developed by its faculty and staff into commercial products and therapies through collaborations with clinical investigators and establishment of corporate alliances. For more information, visit: http://wyss.harvard.edu/.
The Wyss Institute for Biologically Inspired Engineering at Harvard University seeks a talented, enthusiastic and experienced individual to work as a Lead Senior Staff Engineer in the area of bioinspired robotics. This person will report to the Operations Director and will assist and enable Wyss Institute faculty and researchers in their research and development efforts in robotics, while leading efforts within the Institute’s Bionspired Robotics Platform. Topics in bioinspired robotics include, but are not limited to, autonomous robots, collective operation, microrobots, and soft robots. He/she will initiate, direct, and perform independent research and work with Institute faculty in the creation and execution of research programs and projects in bioinspired robotics. The successful candidate will develop technology plans for Wyss robotics research programs, coordinate resources, and champion technology translation.
The Lead Senior Staff Engineer will take a leadership role in developing industrial relationships and outreach to generate feedback that will guide the research and development efforts at the Institute. The successful candidate will participate meaningfully in the academic community, helping to define and create the technology culture at Wyss and serving as a mentor for other senior and junior staff, fellows, postdocs, and students. He/she will participate in the development of strategic plans for Wyss research programs; and in writing research proposals, patent applications, presentations and publications. He/ she will provide guidance and management for prototype development in the bioinspired robotics area and guide applied research primarily focused on achieving commercial proof-of-concept and demonstration of commercial viability. In addition, the Lead Senior Staff Scientist will participate in the selection and management of core and specialized equipment, capabilities and technologies within the Institute to facilitate the Institute’s research and product development efforts. He/she will provide leadership and expertise to research personnel regarding project goals and objectives, as well as Wyss Institute community building efforts around robotics including seminar series and short courses.
Basic Qualifications: PhD in Mechanical or Electrical Engineering or Computer Science or work equivalent with a specific focus on a high risk area of robotics is required. Minimum 7 years of industrial experience in robotics is absolutely necessary. Strong foundation micro and macro fabrication, sensors and actuators, controls (theoretical and applied), materials mechanics, fluids is required.
Additional Qualifications: Strong organizational skills, creativity, motivation, and communication (oral and written) skills, especially ability to work and communicate with non-chemists are required. Experience with grant writing, sponsored research, and project management is essential. Ability to tackle multiple projects simultaneously in a dynamic team environment is essential. Must be able to work in chemical and biological laboratory environments. Ability to assess potential market opportunities and risks and to develop strategic plans for successful technology transfer and commercialization is required. Excellent writing skills and strong commitment to educating and mentoring students and postdoctoral fellows are required.
Interested applicants can please apply directly to our website: www.employment.harvard.edu (please search on Auto Req ID 22597).
February 19, 2010,
Incoming email. The question is longer than the answer.
Q. A few of us at work have been debating a few obstacles in robotic engineering and what is going to be the most difficult to do. We would really appreciate your insight and thoughts on the matter being a reputable Robotics Engineer with a PhD. On one side people think that movement & Hi is going to be the most difficult in creating humanoid robots. They think that dealing with a constantly changing environment, wind, obstacles and getting pushed or bumped even falling over and getting back up is currently the most difficult hurdle to overcome. On the other side people think that being able to visually identify objects the way humans do and to completely understand all of its surrounding and what’s going on will be the most difficult. This would require a robot to not only be able to see and identify objects around him extremely quickly but be able to learn, think and have common sense like a human as well. Since you are a Robotics Engineer we thought you could help us with something that’s been an ongoing debate for quite some time now. We are all very interested in the topic and are excited to hear back from you. Any help you can provide is much appreciated!!
A. Matching the complexity of the human brain is the real challenge. Whether you call it dealing with uncertainty or pattern recognition is just a matter of words. I'm very confident that digital computers will never replicate human understanding. The math is clear. Take a look at a couple of the posts
from November of last year below.
February 6, 2010,
A few questions from some smart eight year olds in East LA.
Let's start with the word “robot.” Karel Capek coined it in his 1921 play Rossum's Universal Robots. The word robota literally means work, labor or serf labor, and figuratively "drudgery" or "hard work" in Czech and many Slavic languages. Traditionally the robota was the work period a serf had to give for his lord, typically 6 months of the year (definition from Wikipedia). These robots were artificial people made in factories. They had no souls and hence could be enslaved without trouble to their creator’s human conscience. Ultimately they rose to destroy their human makers and eventually became extinct because there was no one left to make them. Isaac Asimov’s “I, Robot” series is my favorite exploration of the robots of science fiction. Susan Calvin is the star of the series. She is a human psychologist studying robots.
The definition of robots as manufactured humans persisted until the mid 1950’s when Joseph Engleberger called a machine he was trying to sell into the automobile industry a robot because the name sounded cool. Since then the word has been used to describe just about any computer-controlled, electro-mechanical machine. Note Capek’s robots were biological creations; androids, rather than robots. For now, let’s just call things made by humans that do the kind of work humans do a robot. Hmm, that definition includes my kids. Let’s go with machines that do the kind of work humans do.
Q. In what year will robots be able to talk and look like humans? (I think here we're thinking of androids in science fiction movies and how far off they might be.)
A. It won’t be in our lifetimes, our children’s lifetimes or their children’s lifetimes. The computers we’re using now will never replicate the human mind. Today’s computers are binary in nature, ones and zeros. It doesn’t matter how many mega bytes they have, our brains are infinitely more complex than that.
Q. When will we have robots that can cook?
A. We’ve already got them. Haven’t you ever seen a bread-making machine? What about the instant mac and cheese in the microwave?
Q. When will we have robots that can do our homework?
A. They’re called Mom and Dad. Anyways, would a robot doing your homework help you learn? Maybe.
Q. Why do we make robots?
A. So the robots can do the drudgery and we can spend our time doing more important things; like writing, the arts and watching the
Superbowl!
Q. How long can robots live?
A. By definition robots don’t have souls and are never really alive. Their nuts and bolts can last a few decades before they go into the scrap heap.
Q. What does a robot run on? (e.g. batteries, some other power source?)
A. Robots usually run on electricity. Sometimes gasoline engines power them. Ultimately, just about all of the energy we (humans and robots) use comes from burning things like oil, coal, gas or wood.
November 26, 2009,
Computer simulates cat brain, Not!
A week or two ago IBM announced they had simulated a cat's brain
using one of their super computers. The facts in their announcement
are correct, but the conclusion is completely wrong. The fundamental
mistake was that they simulated a neuron as a single point digital
connection. The connections in a mammalian brain are far more
complex than that. Each connection involves ion channels, multiple
branches, electro-chemical interactions, and on and on. To quote Henry
Markram's (the lead on the EPFL Blue Brain project) opinion on IBM's
simulation, "These
are point neurons (missing 99.999% of the brain; no branches; no
detailed ion channels; the simplest possible equation you can
imagine to simulate a neuron, totally trivial synapses; and using
the STDP learning rule I discovered in this way is also is a joke).
He wrote this in an open letter
to the respected IEEE. I encourage you to read his letter. You can
also take a look at my post below. To reiterate, there is no path
from digital computers to the complexity of the human brain.
November 10, 2009,
How long before Star
Wars robots?
Here's a rare article that talks about how far away we are from having robots with human-like capabilities, rather than how close we are to that day. I definitely agree with the article. I don’t think we’ll see robots with anything like human reasoning this century or that we’ll necessarily ever see it. People who do say we’ll have robots 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 by Intel co-founder, Gordon Moore. Second, the interconnections in human brains are much more complex than binary computer connections. The connections in human brains 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 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.
June 10, 2009, RoboGames is only a few days away!
This Friday, Saturday, and Sunday, the world's largest robot show returns
to San Francisco! Hundreds of robots from around the world competing in
over 50 different events. Giant fighting robots, agile acrobatic androids,
intelligent Legos, soccer bots, and much, much more:
"Top 10 Video Highlights" - ESPN SportsCenter
"The Best Ten North American Geek Fests" - Wired
"The best robots compete in RoboGames, just as the best athletes train for
the Olympics." - Discover
"World’s Largest Robot Competition" - Guinness Book of Records
Still cheaper than a Giants game, and hey - we got androids, not steroids!
Friday, Saturday, or Sunday - $20 a day or $55 for all three. Discounted
tickets for kids, as always. Buy your tickets today!
Fri-Sun, June 12-14, 2009 at Ft. Mason Festival Pavilion.
$20/adult, $15/kids
PARTICPATE IN THE ACTION!
How would you like to build a robot with your kids? The new Robotis Ollo
kits let you do just that! Each day, Robotis will be running 3 workshops
to teach you how to build small robots and compete them! There are 3
workshops each day, with only a limited number of people per workshop.
Each team is good for one adult and one child.
Ollo Action kits are only $29.95, including the workshop and competition.
Sessions are at 12 and 2pm each day.
Ollo Bug kits are $99 which includes the workshop, two free drinks, and the
competition. Sessions are at 4pm each day.
To reserve a space for the Ollo workshop, and build your own robot that you
get to take home, please email dave@robogames.net Please include which day
and session you're interested in.
DONATE
RoboGames is still an all-volunteer effort. We need your generosity in
order to continue. We realize that the current economy sucks, but anything
you can contribute will help to make this year great, and next year even
better! If you can't donate - email us to find out how you can volunteer
and help out - we need people both at the event, and leading up to it (as
in: now...)
May 25, 2009, More government
funding for robotics?
Rep. Phil Gingrey (R-Ga.) is looking to increase government
funding for robotics.
http://www.ajc.com/services/content/printedition/2009/05/25/robot0525.html
Of course I'm a supporter of robotics, but this spending is
getting out of hand. Printing infinite money will eventually catch
up with us.
May 20, 2009, Robots in movies
sampler
From Marie in the 1927 silent movie classic "Metropolis" to the Terminator franchise,
has a nice overview of movie robots through the years.
January 24, 2009,
"Glazed" a painting by Eric Joyner
Take a look at Eric's website and
consider buying some prints or a book or something else cool there.
December 13, 2008, Dancing Robots
Robots in the new Motoman Robotics Lab gave a "command" performance Nov. 21 to celebrate their new home.
Dancing robots, directed by a master robot, performed to the "Dance of the Sugar Plum Fairies," the opening to Star Wars and "Mr. Roboto" to celebrate the dedication of the Motoman Robotics Lab at the University of Dayton, Friday, Nov. 21.
In addition to enhancing the arts scene, the new robots will provide academic and research benefits to Ohio. The new lab could enhance Ohio manufacturing capabilities and provide opportunities for Ohio industry to partner with UD to research new ideas.
Two Ohio companies — SAS Automation and Motoman — already have made investments and contributions to the University valued by the companies at more than $500,000.
UD President Daniel J. Curran, Joseph Saliba, interim UD provost; Don Moon, chair of UD's electrical and computer engineering department; Steve Barhorst, Motoman president; and Trent Fisher, SAS Automation president, spoke during the dedication in Kettering Labs Room 232.
April 12, 2008, Terminator Series
Review
First of all, I give a "two thumbs up" to the
Terminator TV series. That's no small endorsement as I have very
high expectations for the Terminator line (including the movies).
The time travel stuff is always fun, but of course it's the robotics
that I like. Unfortunately, we know what it is like when humans lack
morality. This series explores super-humans
with no morality. To all future roboticists: Please program Asimov's
three laws into your creations.
Jan 25, 2008, Robot Learning
As mentioned below, only a fifth of the cost of deploying a robot
on a factory floor is the cost of the robot itself. The rest of the
cost is programming the robot, developing the tools and developing
the manufacturing processes. A company called Skilligent is working
to reduce the cost of programming the robot. They announced the commercial release of
a robot learning software. The main concept behind the software is that it can learn skills directly from humans without programming. This opens the door for building robots
that don’t require traditional programming as they can be trained “out-of-the-box” by a non-professional user.
Jan 3, 2008, Indian robots toil on global shop floors
Lower labor cost is the reason often cited for off-shoring
manufacturing jobs. So why would it be cheaper to use robots in
India than in the USA? It's because only a fifth of the cost of
deploying the robot is the cost of the robot itself. The rest of the
cost is programming the robot, developing the tools and developing
the manufacturing processes. It will be a long time before we
automate that part of the equation.
|
|
