A Final Mission For NASA’s Discovery Shuttle Marks A New Beginning For Humanoid Robotics

February 26th, 2011

This past Thursday, February 24th 2011, marked two distinct milestones for NASA and space exploration: space shuttle Discovery launched for a final time bringing its 27 years and 143 million miles of service to an end while at the same time carrying in her cargo bay Robonaut 2, the first humanoid robot (that we know of) to be deployed into outer space.

Robonaut 2 Being Carried on Space Shuttle Discovery STS-133

Robonaut 2 Being Carried on Space Shuttle Discovery STS-133

To be more specific, Robonaut 2 is only half a humanoid robot, containing an upper torso and a head but lacking hips, legs, or feet… NASA plans to add those in later versions. R2, as the robot is called, has been developed in collaboration with General Motors (the automotive company), and is what NASA expects to be the first of a new generation of robots that will assist humans in space explorations. As such, the deployment of R2 at the ISS will comprise the first wave of various experiments in deploying this type of robots into outer space, tasked with both assisting and collaborating humans in their missions while also carrying out missions that are otherwise too dangerous or laborious for human astronauts. Among these tasks, EVAs, Space Assembly, and General Repairs of ISS’ outer structure are amongst the most cited examples of what they expect to do with R2.

R2 is capable of both teleoperated missions as well as autonomous and semi-autonomous modes. Certainly having dextrous hands capable of using some of the same tools as humans do will go a long way in carrying out these missions and NASA expects to both accomplish and learn a great deals in using robots in such missions in outer space.

A while ago I had a brief email thread with Dr. Firouz Naderi at JPL about the wisdom of the humanoid form-factor for such robots. While the crammed hallways of the ISS are conducive to the humanoid form, I can think of a variety of other form-factors that may be more ideal to the Zero G environment and the types of tasks they are expected to perform – particularly in the longer term roadmap of NASA’s space exploration roadmap. Nevertheless, the fact that NASA is actively investigating the role of robots in space exploration marks a significant milestone not only for NASA but for robotics as well.

No doubt we will be learning a great deal on human-robot interaction given the tightly controlled and scripted experiments that are run on any given day on the ISS. The introduction of robot assistants into the mechanics of running these experiments will go a long way in providing best practices and intuitive interfaces for human-robot collaboration, amongst other lessons. The use of robots also opens up the possibility of experiments that are otherwise too dangerous and/or laborious for human scientists to undertake in the extreme environment of LEO, leading to both direct and indirect scientific research benefits never before possible.

There are other benefits worth mentioning as well.

As robots are a natural at capturing the cultural imagination, putting a visible humanoid robot into space may have the added benefit of getting people interested in not only space exploration, but the incredible scientific research and development that is needed and conducted on an ongoing basis to do so. The interest of the populace is particularly important today as the global economic climate and the domestic political mood are in misalignment with our appetite to invest in such endeavors.

It is indeed worrisome that the current Republican controlled U.S. Congress has publicly gone on the record that cutting funding for NASA is on their agenda. While as a CEO of a for-profit company I am deeply concerned with our national debt, I believe cutting funding to NASA (and any other national organization conducting public R&D for that matter) is extremely short-sighted and carries grave consequences for the longer term viability of our national economy. It is no secret that innovation has been the key driving force of our economy and if we are to remain competitive in the globalized marketplace, it is vitally important that we continue to be a nation of innovation. And without public investments in large-scale abstract and directed research, we will be severely and measurably cutting the very lifeblood that has sustained our economic viability since World War II.

R2 is a great first step… and to the extent it helps to inspire us to ever greater achievements, and inspires an entirely new generation of our youth into science and engineering, I’m all for it, wish it luck, and hope to see it do great things in the very near future.

Space Shuttle Discovery STS-133 Crew

Space Shuttle Discovery STS-133 Crew

So bon voyage to R2, Discovery, and her incredible crew of astronaut scientist.

Best,
/F

Space Shuttle Discovery STS-133 Launches for her Final Journey

Stunning video captured on iPhone from airplane



The Need for Low Cost Sensors in Robotics – Holiday Edition

January 12th, 2011

Microsoft Kinect Shows the Robotics Industry a Simple Roadmap for Developing and Commercializing Complex yet Low Cost Sensing Using Commodity Components and Constantly Declining Cost of ComputationTravis Deyle over at Hizook asked me to write an article for his blog which I was happy to oblige given his dedication to sparking intelligent and in-depth conversations on topics in robotics.

The Need for Low Cost Sensors in Robotics – The Holiday Edition” is an analysis of the current state of sensory components in robotics, our dire need for more complex sensing at lower prices, and a look at how Microsoft has been able to create a powerful (3D sensing) device with Kinect at a very reasonable price.

Emulating the same strategy as Microsoft, the robotics industry can also achieve significant milestones in commercialization of mobile/intelligent robots… and the Kinect project should be looked at as a simple and achievable roadmap in our industry.

Chris Anderson, Wired Magazine’s Editor in Chief and author of one of my favorite books The Long Tail, seems to agree with me on the tenets of my analysis. On his DYI Drones blog, Chris writes about the article:

    There’s a very good post over at Hizook on the lessons of the hacked Kinect: Nutshell:

    “The best solution to complex low cost sensing (or actuation for that matter) is to take advantage of affordable, mass-produced components, complementing them with the innovative use of software solutions that benefit from constantly declining prices of computation.”

    Well worth reading it all.

You can read the full article over at the Hizook blog here:

http://www.hizook.com/blog/2011/01/09/need-low-cost-sensors-robotics-holiday-edition/

Special thanks to Travis Deyle for posting it to Hizook.

Kinect by Microsoft XBox 360 gaming platform is a powerful 3D image and depth sensor that has applications far beyond gaming, particularly in roboticsGiven the feedback I’ve received on this article, I’ve been inspired to write a whole series of articles analyzing and offering a roadmap for the successful commercialization of personal robots. If you have insights and would like to contribute, please get in contact with me.

Happy new year everyone.

Best,
/F



200 Years of Global Life Expectancy in Four Minutes (Dr. Hans Rosling)

December 13th, 2010

Large data-sets have always fascinated me.

Almost all countries in the world have been keeping reliable statistics on things like life-expectancy and GDP for over two centuries. The advent of digital technology and personal computers specifically has enabled us to not only collect better stats on everything, but to analyze the numbers and gain real intelligence out of them.

This analyzing of the date to gather intelligence is what specifically interests me. In more recent years, as computer generated graphics have excelled, so has our ability to display data in ever more creative ways. Technologies such as Augmented Reality, the notion of transposing reality with a layer of computer-generated graphics, is a very effective means for data visualization. AR has the potential to enables the layman to understand, analyze, and appreciate the massive amounts of data we collect on everything.

Case in point: Dr. Hans Rosling and Global Health!

Dr. Rosling is a distinguished Professor of International Health at Karolinska Institute and Director of the Gapminder Foundation. His work often involves looking at vast amounts of data to better understand the macro trends of Global Health, enabling him to make better predictions about our future and informing the tools and infrastructure that we’ll need going forward. In this short and fascinating video produced by BBC, Dr. Rosling visualizes 200 years of life expectancy at a global scale, in under four minutes with a little help from Augmented Reality.

As a side note, the Bill and Melinda Gates Foundation which is the largest philanthropic organization in the world, has identified Global Health as its international objective. Their mission is to eradicate the 10 most deadly diseases worldwide through research, philanthropy, and education.



Gartner Gives Love to Robotic Telepresence

November 22nd, 2010

Gartner’s Jackie Fenn has published Executive Report called Six New Technologies That Will Reshape Your Business. You can read the summary of it here (registration required).

Gartner LogoThe six technologies identified are:

  • The real-world Web of smart objects and places
  • Augmented reality, which provides contextual information as an overlay on the physical world
  • Next-generation user interfaces, including touch, gesture and tangible computing
  • Predictive analytics of new data sources and types
  • 3D printing of physical models and spare parts
  • Mobile robots for remote presence and mobile infrastructure

I’m of course very glad that serious research houses are giving Robotic Telepresence their due respect, and finally distinguishing the difference between traditional HD Conferencing and Mobile Robotic Telepresence correctly.



The Need For Low Cost Sensors

November 19th, 2010

Microsoft Xbox Kinect sensor used on a mobile robot platform - Philip Robbel at MITWhen it comes to sensors and robotics, the prevailing school of thought is let’s create more complex sensors and drive the cost down. Complex sensors enable us to make complex models of the world, allowing robots to make sense of the worlds they live in at a granular and nuanced level. The tradeoff is that more complex sensors cost more money, and until/unless someone drives down the cost through scale (and innovation), they’re cost-prohibitive for consumer level applications. But on occasion, someone does find the incentive to use innovation and scale to drive down the cost of a complex sensor, and the results are impressive.

Case in point: here’s Philipp Robbel, a PhD candidate at MIT, hacking a Microsoft Xbox Kinect sensor and mounting it on a mobile platform for 3D SLAM… and as a bonus, gesture recognized commands.

As Robel puts it, this is a “weekend hacking project” that any brave soul (with the appropriate skill set) can do. The results speak for themselves: a relatively inexpensive platform that can do some impressive and complex SLAM.

Here’s what’s interesting to me though: the reason Robel can do this has nothing to do with robotics. It has to do with gaming. Microsoft, in an arms race against Nintendo Wii to find a better control mechanism for gaming, went out and found a creative solution involving sensors that would otherwise be still made in low quantities (and high prices).

Cross-functionality is an inherent beauty of technology. The things we discover and develop in one context has application in other domains. Gaming has been a impressive driver of digital technologies for decades. Graphics processors (GPU), sensors, and storage devices are just a few examples.

At the same time that I argue for driving down prices of complex sensors however, I’d like to put forth another approach to robotics that has long been forgotten by our university and institutional robotics research. That is, the approach of using more of existing low cost sensors on the market. In many cases, a better (and more economical) approach to sensing is not just more complex sensors, but using the dumb sensors we have in larger numbers. We have advance algorithmic and filtering techniques today that can do a great deal with the data streams that existing low cost “dumb” sensors can produce. This approach has been universally abandoned in favor of using expensive complex sensors – and it amounts to a disservice to the types of robots we can develop.

Certainly we should put resources into developing complex sensors, and exploring market-driven approaches to drive costs down. But getting robots into every home, where cost constraints are unforgiving, requires of us creative approaches that amongst other things, include using what we already have (inexpensively) to make sense of the worlds our robots live in.

I will write in more detail about this topic in the coming months, as I’m turning my attention to making robotics more accessible to wider audiences. In the meantime, I wanted to write this post to start a conversation amongst the people doing this type of research.

I look forward to your thoughts and feedback.

Best,
/Fred


Resources

  • Phillip Robbel – http://robotic.media.mit.edu/people/philipp/philipp.html
  • Open SLAM – http://openslam.org
  • Xbox Kinnect – http://www.xbox.com/en-US/kinect


    Update:
    In a highly disingenious move by Microsoft, spokespeople for the company now say that Kinect hardware was “left open by design” to encourage hacking by users to find new uses for the device. This is despite earlier posturing by Microsoft, in response to a $3,000 bounty for the first people to come up with an open source driver for Kinect, that “…Microsoft does not condone such behavior” and that “we will work closely with law enforcement to discourage such behavior”. Adafruit Industries who put up the bounty wrote in a new blog post that “In about one week we turned [Microsoft's contention that it would stop hacking of the Kinect by working] ‘closely with law enforcement’ to ‘inspired’ by community finding new uses for Kinect.”

    The good news is that at least now the company supports “new uses” for Kinect and perhaps we can see prices (and usage) drop down further as we go down this path.