Bringing Research into Operations in NASA's Mission Control

Many companies in the world have tried to do Research and Development (R&D), but have failed to do this effectively. There are few companies left who have the budgets to keep a R&D facilities (Microsoft, IBM, HP, to name a few). Look what happened to Xerox Parc, Xerox' golden R&D facility in the eighties. Look at Bell Labs, Ma Bell's R&D facility since the twenties. Look at Bellcore, the Baby Bell's R&D facility, after the breakup of Ma Bell. And, look at NYNEX Science & Technology, the R&D facility of the former NYNEX Baby Bell company (now Verizon), where we started working on the Brahms Multi-Agent Modeling and Simulation environment in the early nineties. Gone are they!

Companies create a R&D center, let researchers go at it for a while, after which they realize that researchers are expensive and don't create many useful products for the company. Consequently, the R&D is stopped, because "outsourcing" R&D seems to be more lucrative.

NASA has an opportunity to do it different. NASA Ames Research Center is truly one of the best places where a computer scientist, interested in Space Exploration, can work. The thrill is to be able to do both basic research when needed, but then to move your research into mission operations as well. In the first seven years being at NASA Ames (1998 - 2005) we received plenty of research funding to develop our Brahms environment. I finished my Ph.D, we did many Planetary Exploration Expeditions to Haughton Crater on Devon Island in the High Arctic, Meteor Crater in the Arizona desert, and the Four Corner region, near Hanksville in Utah (I have the T-shirt that says "Where the hell is Hanksville?"). We were able to develop Brahms as both a simulation environment, testing it out on the Mars Exploration Rover mission at JPL, as well as a real-time Agent-Oriented Language to develop distributed multi-agent systems (MAS) for planetary exploration.

We integrated our agents with speech dialogue, bio-sensors, GPS, digital mapping, semantic web databases, wide-area wireless networks over many kilometers in the desert having agents sent e-mails, digital voicenotes, pictures downloaded from normal digital cameras, and panoramas from robot cameras to people in Buffalo, New York, San Diego, CA, Milton Keynes, UK, and Canberra, Australia. It was cool work, new, innovative. Way before Google existed, let alone was able to think about Google Maps or Google Earth, we were already tracking people and robots, linking images and digital voice notes, all real time, via GPS in the Utah and Arizona desert. Basic and applied research is what we did, over a seven year period. Very cool and a blast to do.

Then, Mike Griffin became NASA's new administrator. My $15 million research project got cancelled after only 6 months of the 4 years. Devastated we were. NASA had to get a new focus, a real objective to work towards. Mars was out, the Moon was in. Back to the Moon we will go, Griffin said. Humans and robots exploring Mars was a topic we weren't allowed to work on anymore. Ok, so now what? Science is out, engineering is in!!

Our first reaction, as is still the issue for the hard sciences at NASA (planetary science, Earth science, Astrobiology, etc), was that we are researchers, not engineers. Wrong we were. You had two choices; 1) Leave, which a lot of our computer science colleagues did. To Google, next door, they went. Funny, although they didn't want to develop tools for Space Missions at NASA, they now have titles like "Quality of Search." They are doing cool work at Google, don't get me wrong, but researchers they are no more; 2) You can stay and see what this new mission brings.

Staying, we did, and with some success, although it didn't come easy. We lost our research funding and had to go and "sell" our technology to the operations centers (Johnson Space Center, Kennedy Space Center, and Jet Propulsion Lab). Successful the TI division at Ames has been in doing this. In our case, we are now getting very close of having our Brahms environment automate some of the flight controller's work process at JSC's Mission Control (you know, from "Houston we have a problem!"). Sixteen years we have worked on Brahms, starting in 1992 at NYNEX Science & Technology. From basic research (I even was able to write my doctoral thesis on Brahms) to now automating parts of NASA's Mission Control.

As I tell my friends, how many people can say that they have helped develop systems for NASA's Mission Control, let alone, how many researchers can say they have done research for 15 years and now their work helps NASA to do space missions?

It is but a small step for mankind, but a huge victory for the Brahms team :-)

To everyone who reads this and who has worked on Brahms at one point or another (you know who you are), we are planning to give a huge party after Brahms goes live in Mission Control. When this is is anybody's guess at the moment of writing this. The plan is before end of July this year (2008), but many software control boards have to be passed, before we are able to claim victory. I will keep you all posted.

I will end by saying that I think NASA has the opportunity here to get the benefits of years of research that they funded. I am convinced that successful R&D can be done by funding research in phases, and then switch to fund bringing the research into operations. This is true R&D. Seven research years, followed by two to four development years. Then, go back to research. Back and forth. NASA can show industry how it is done ... I'll be even more stronger in saying that NASA *is* showing industry how it should be done.

The American tax payers should demand that Congress stops funding useless Pentagon projects, and give some of that wasted money to NASA. This, my friends will give a return on investment that the future generations can be proud off.

On to the Moon we will go, and beyond ...