Linux camera launched at Lunar X PRIZE
Dec 17, 2008 — by LinuxDevices Staff — from the LinuxDevices Archive — 6 views
Foreword — This article describes an all-volunteer science project's use of a donated Linux camera for the Jaluro rover it hopes to field in Google's Lunar X PRIZE competition. It discusses several advantages that patent-free Open Source software like Linux and Theora offer for scientific research projects. Enjoy . . . !
by Anders Feder
Introduction
Have your eyes ever wandered across the crisp scans of some of the millions of pages recorded in the virtual Google Books library? Have you ever gazed at the detailed sceneries in photos from cities all over the world through the Google Street View service? If so, you might be surprised to learn that the next time you'll come across a view like that, it may not be set in the streets of Paris or Seattle, or any other location on Earth for that matter, but rather on the surface of the Earth's distant natural satellite, the Moon.
The cameras used by the two aforementioned services, to acquire images of book pages and street views respectively, were produced for Google by a small Utah-based company called Elphel, Inc. The company manufactures so-called digital network cameras — cameras that serve images and video to a network, such as the Internet, rather than to a single computer like standard digital cameras do.
Elphel is far from alone in the market for network cameras, but unlike products from imaging and optics behemoths like Kodak and Canon, Elphel's cameras are uniquely developed and delivered on the same non-proprietary Open Source terms that software projects like Linux and Firefox are best known for. Each physical camera sports a discrete, punning “Source Inside” logo on its exterior, and comes with the complete source code of Elphel's software, plus links to the repositories of large third party projects, such as the Linux kernel, PHP, and others. If the customer installs the source code on a Linux computer, a fully functional firmware “image” can be built from source, by running a single script. The new firmware image, with any adjustments the customer has made, may then be installed in the camera.
And as it happens, another Open Source project, Team FREDNET, is currently designing a remote-controlled robotic rover intended to be sent to the surface of the Moon, and they were coming up short in the search for a digital camera suited for returning images of the lunar landscape, as stipulated by the rules of the contest in which the team participates, the Google Lunar X PRIZE (of which Elphel's aforementioned customer, internet search giant Google, incidentally is main sponsor). The project then discovered that one of Elphel's cameras matched the requirements perfectly, and a few weeks ago, the two projects could announce that they had, in the classic collaborative style common to Open Source, formed a strategic partnership under which Elphel will provide the main camera for Team FREDNET's lunar rover.
Team FREDNET's contact with Elphel was established on the initiative of Tobias Krieger, an Austrian software and electrical engineer working at Austrian Airlines. Krieger leads one of the groups within Team FREDNET that is developing a rover concept he calls “Jaluro” (short for “Just Another Lunar Rover”). Jaluro is a minimalistic vehicle with just two wheels.
The complete Jaluro rover, with camera
(Source: FREDNET. Click to enlarge)
Krieger began work on the concept of a two-wheeled vehicle because he missed variety in the rover design process, and to incite others to think creatively about rovers as well. Eventually, these efforts led him to the rover's visual system, a particularly challenging part of the design. Krieger explains, “Independent of the rover design, we have to fulfill several tasks according the Google Lunar X PRIZE guidelines. One of the major tasks is the visualization of our trip on the lunar surface, with very high requirements on the imaging technology.”
Current draft rules from the X PRIZE Foundation requires that if they want to be eligible for the competition's $20 million grand prize, teams must return high-resolution, low-noise video, in a quality comparable to home cinema. Says Krieger, “Cameras that can deliver the required quality are a very advanced piece of technology. Team FREDNET is, by virtue of the weight and size exigencies of a lunar mission, very limited with respect to the components we can pack into the rover. We don't have much power available, hence only low-power equipment can be used,” he adds.
Once on the lunar surface, the team's rover will use a wireless connection back to the lander module that deployed the rover, and the lander module will then serve as a relay, re-transmitting data it receives from the rover back to the Earth at high power. To waste minimal power beaming the bandwidth-intensive high-quality video data from the rover to the lander, the signal must be compressed prior to being transmitted from the rover. “That leads us to a requirement to do real-time HD video and image encoding and compression right in the rover — with small dimensions and almost no power consumption,” Krieger says.
And that's where Elphel comes into the picture. Acting as embedded web servers, Elphel's network cameras are small computers in their own right, more than capable of accomplishing the encoding and compression tasks to be done by Team FREDNET's rover. In fact, the cameras — which run commodity Open Source server software such as Linux and PHP — are so advanced that Krieger is even considering using them for general computing jobs such as controlling the rover's wireless network connection to the lander.
A further advantage, of course, is Elphel's distinct licensing terms, which are similar, if not identical, to those of Team FREDNET. “Elphel supports Open Source codecs, and Elphel technology is itself Open Source,” Krieger states. “In my opinion, an Open Source project like Elphel, and Theora, the codec they use, are the best fit to an open space project such as Team FREDNET. Their technology is no secret — there's no non-disclosure agreement (NDA) to sign — it's an open community, and the camera development is well-maintained,” he says.
But to Krieger, perhaps the most important selling point was the company's willingness to cooperate: “I got such a good feedback on the first e-mail I wrote to them that I knew they would be a good choice.”
Elphel backgrounder
With seven years in the business, Elphel is truly an Open Source company. It was founded by Andrey Filippov, a Ph.D. graduate from the Moscow Institute for Physics and Technology born in Russia. In 1995, Filippov moved to the United States to work for an imaging company in Salt Lake City, Utah, but after six years as a payroll worker, a new development model evoked Filippov's entrepreneurial spirit. He recalls, “When I learned about GNU/Linux and noticed how efficient it is for the development compared to the proprietary solutions, I just quit my job, and founded Elphel. Our company was established in 2001.”
Today, Elphel offers several advanced network camera models with fully open schematics and source code. “All our business is 100 percent based on GNU GPL — we never agreed to license our development on any other terms, although we had such offers,” Filippov says. “I just believe this business model is viable, and it has worked fine for us so far.”
Alexandre Poltorak, who is in charge of developing Elphel's business model and relations to the Open Source community, notes that while outside contribution to Elphel's own work is not overwhelming, as much as 97 percent of the firmware used in the company's cameras stem from popular Open Source projects, such as Linux and the GNU C Library, which have huge, vibrant communities of their own. Elphel, in turn, has contributed updates and hardware to other Open Source projects, including the Axis Linux distribution, and MPlayer, a media player. Poltorak explains that it is the company's policy to donate one camera per month to projects developing Free Software that may benefit Elphel and its users.
Scientific benefits of Open Source
Tobias Krieger says that the availability of Open Source technologies such as Elphel's cameras are instrumental in keeping his team's lunar mission an open effort. “Nowadays, many sophisticated technologies, both hardware and software, are protected by an immense amount of patents,” he says. “For small projects like ours, it is very hard not to step into a hidden trap. Most manufacturers require their customers to sign an NDA before sending samples and data sheets.”
The manufacturers use NDAs to keep customers from passing on details about a given technology to people outside their organization, in an attempt to keep the technology a secret and capitalize on the knowledge used in its creation for as long as possible. But in the case of Team FREDNET, the secretive agreements also keeps potential outside contributors from evaluating the team's systems and suggesting improvements, and that's a problem, since outside contributors constitute the team's main source of labor. “Open source, as Team FREDNET is, and an NDA do not agree very well,” Krieger says. “My personal opinion is that patents and secrets slow the evolution of technology down.”
Open source partnerships — built on trust
Proprietary or not, Elphel's cameras are still very sophisticated pieces of hardware that require a good deal of resources to produce, in terms of know-how, manpower, and materials. Without NDAs, or even receiving money in return for the hardware, is it really a smart business move to hand over such an expensive piece of equipment to a group of people literally intending to blast it into outer space at their first opportunity?
The answer is that it might be. Elphel intends to evaluate Team FREDNET's progress as the mission develops, and gradually decide to what extent they wish to participate. “The partnership between Elphel and Team FREDNET isn't based on contracts, it's based on confidence,” Krieger says. “We both have to trust each other to comply with the arrangements. A good question is what that trust is based on. I think we have to go step by step, and finally we will see that we can trust each other.”
For now, Filippov has agreed to donate an Elphel NC353L-10369 camera to Tobias Krieger's team, free of charge. (The camera is similar to the NC353L) Krieger will then integrate the camera in his lab, testing how well it works together with a live streaming server, and what data transfer rates can be achieved when using the camera with a wireless connection.
Elphel NC353 camera (note “Source Inside” logo)
(Click to enlarge)
Together with Jörg Schnyder, another rover designer, Krieger will develop a print board with exactly the features the team need, including the capabilities to control the camera, and to route the encoded and compressed image data to the wireless connection. Filippov says his company does not have the time to participate directly in the development of the visual system at the moment, but Krieger's team will be able to seek him out for advice and information during the integration process.
Incidentally, apart from being the lead designer of Elphel's cameras, Filippov also once worked at the institute which developed a laser for Phobos 2, a Soviet space probe launched in 1989, destined for the eponymous Martian moon. The laser was to act as part of a spectroscopic device, vaporizing a small piece of rock on the moon, and analyzing the light emitted during the burn to determine its chemical composition, similar to the “ChemCam” being included on NASA's next Mars rover, the Mars Science Laboratory. Like many other Martian probes, Phobos 2 failed in its final phases, and regrettably never got the chance to fire the laser which Filippov's institute developed, but Filippov is now trying to get in touch with his old contacts from the project to see if they can help out with Team FREDNET's mission to the Moon. Such expertise would be greatly appreciated on the team, as Elphel's otherwise consumer-grade cameras face a tough challenge in the harsh space environment of cosmic radiation, extreme temperatures, and hard vacuum.
In return for their generous sponsorship, Elphel is offered a number of benefits by Team FREDNET, including exposure through the team's website, and potentially even in the video transmitted from the Moon, as well as early opportunities to engage in media events, demonstrations, and so on. But Krieger says Elphel has been positive from the outset, and that it didn't take more than two e-mails to get the company on board. “It is just an exciting project, and can put our products to a real test,” Filippov says.
“It can also give us some good publicity if our common approach of Open Source wins that challenging competition,” Filippov adds, referring to the Google Lunar X PRIZE, in which 16 other skilled teams participate — some with much greater resources than the small non-profit Team FREDNET. Filippov isn't in doubt about sponsoring a non-profit team, however, stating, “We would not sponsor a proprietary mission. Sell them our products, sure, but why sponsor?”
“'Free software' and 'Open Source' are not just commercially attractive words for us, it's our business model,” Alexandre Poltorak says. “Following this business model, even if we do not land on the Moon, just talking with creative minds, and having an open technical discussion can in itself be very beneficial. Following the same model, but collaborating with and sponsoring a closed source project is totally useless, and may even lead to bad situation with respect to patent law,” he says, referring to how Open Source and closed source are often a bad legal cocktail, because of the contradictory claims to intellectual property the two distinct licensing paradigms make.
Project status and conclusion
Tobias Krieger hopes to have Elphel's camera set up for testing in a few weeks. Integration with the rest of his rover will then commence around next spring, and Krieger hopes to have a finished rover prototype by the end of 2009. Meanwhile, the rest of his team is busy developing the many other components needed for their lunar mission, including the lunar transit vehicle that will carry the rover from the Earth to the lunar surface. No dates for the completion or launch of the whole stack are available yet, but the grand prize of the Google Lunar X PRIZE contest expires in 2012.
Complete details on the rules and schedules of the competition, along with information on the other teams participating, are available from the official Google Lunar X PRIZE website, here.
About the author — Anders Feder is a student of Computer Science at the University of Copenhagen in Denmark. He's been a user of Linux since around 2003, and has been using various open source programming languages, including PHP and C, since around 2000. A few months ago, he discovered Team FREDNET at the Google Lunar X PRIZE site, and opened a blog to report on their progress. Eventually, he decided to join the team, and is now involved at several fronts, including outreach and development of mission control software.
Papers by Andrey Filippov:
- Linux video camera geo-tags, writes to SATA drives
- Open source-based high-resolution cameras for Web developers
- AJAX, LAMP, and liveDVD for a Linux-based camera
- Building an Ogg Theora camera using an FPGA and embedded Linux
- How to use free software in FPGA embedded designs
- Using Embedded Linux in a reconfigurable high-res network camera
- Using Embedded Linux in a High Speed Gated Intensified Camera
This article was originally published on LinuxDevices.com and has been donated to the open source community by QuinStreet Inc. Please visit LinuxToday.com for up-to-date news and articles about Linux and open source.