Kelso, T.S., Keynote address, presented at SMOPS-2023: International Conference on Spacecraft Mission Operations, Bengaluru, India, 2023 June 8.
Namaste, aloha awakea kākou, and good afternoon.
Mahalo and thank you to Chairman Somanath and ISRO, Director Ramakrishna and ISTRAC, IAA, and particularly Deputy Director Nandini Harinath, Amit Singh, and all the conference organizers who did so much to make it possible for me to be here with you today.
I am honored to have this opportunity to talk story with you, as we say in Hawaiʻi, and share the journey of how CelesTrak came to be and all the possibilities that lie ahead on our journey together.
Sometimes it seems as if this story started a long time ago in a galaxy far away. I wish that all of you should live so long to one day understand what I mean by that.
So, how did this journey begin? I guess it started for me by always being curious and was stimulated by growing up in South Florida during NASA’s Gemini and Apollo programs. As a kid, I especially loved astronomy and I got my first real job working at a grocery store just to make money to buy my first telescope.
I carried that passion for learning off to college at the US Air Force Academy, where I majored in Physics and Math. That educational foundation made it possible to understand so much about the universe around us. Back in the early 1970s, though, much of what I learned was just the theory and there seemed little opportunity to put that into practice. After all, there was only so much you could do with the slide rules we were issued when we started!
By the time I graduated, though, we had scientific calculators and I was determined to see what I could do with those. But at my first USAF assignment at Whiteman AFB, MO, I learned I could sign up for a masters degree in business administration and have access to a mainframe computer. It was all punch cards and paper printouts, but I set out to see if I could predict the positions of the planets. Sure, that had been done before, but not by me!
And that was where I got introduced to my first personal computer in 1977. Wow! I could use that at home, not have to carry around decks of punch cards, and get the results right away instead of waiting for them to be sent from the main campus to the line printer. I had to have one and get those planet predictions running on it!
Then there was one of those moments of serendipity that can change a person’s entire life. One evening I was watching the ‘local’ weather on a small black & white TV with a screen about the size of my hand. The TV station was over 100 kilometers away and the signal was never great, but on this evening, I heard the local weatherman start talking about being able to see a dot of light crossing the sky the next night. That dot of light was Skylab, which was about to fall from orbit.
I called my friend and we got together before the appointed time and lo and behold, right on schedule, there was a bright dot slowly moving across the sky. Fortunately, the weather was clear or I probably would not be standing here today. But that was the moment when I thought: I want to know how to do that! And maybe I could do that on my personal computer.
I knew the basic orbital theory from a course at the Academy, but where did the data come from for Skylab? There was no public Internet and no Google search yet, so it was going to take a couple of years to pick up the next piece. But the seed for the challenge had been planted.
Back then, if you wanted to learn anything about computers, you got that from magazines, so I subscribed to quite a few. It turned out there were people trying to apply personal computers to astronomy, but nobody seemed aware of how to track satellites. Then I found an ad in one of the magazines selling some satellite tracking software and it would run on my TRS-80 personal computer. Things finally came together where I could get a copy of the software and then the next phase of the journey began in earnest.
And, soon I was off working on my master of science degree in space operations, with a specialty in astrodynamics, at the Air Force Institute of Technology (AFIT). I pored over the BASIC source code trying to figure out how it worked and what some of the numbers represented. One day, I asked my class leader—who had worked in Cheyenne Mountain—what he thought. He looked at it and said: Oh, that’s SGP4. I remember thinking, what the heck is that?
Luckily, through our US government systems, I was able to get answers to that question, including a copy of Spacetrack Report Number 3. It was all quite fascinating!
But you remember the key challenge was: Where do you get the orbital data? The software package explained you had to write a letter—yes, on a piece of paper, put it in an envelope, put a stamp on it, send it off, and wait a couple of weeks—to NASA/GSFC requesting data for the satellites you wanted to track. But who knew anything about what satellites were up there? I asked for data for a couple satellites suggested in the software documentation and in a couple of weeks I started getting envelopes in the mail with pieces of paper. At the top of these NASA Prediction Bulletins were what we now call two-line element sets or TLEs. By the way, I have a copy of one on my iPad, if you are curious to see what they looked like.
Once you got these documents, of course, you had to type the data into the computer before you could use it. It was easy to make mistakes, so I wrote my own editor to minimize the chance of making those mistakes and it even calculated the checksums. A couple of times a week I would sit down and enter the data so I could predict satellite passes and actually see the results. I was finally able to put it all together and see my first pass of SEASAT over Dayton, Ohio in 1981!
Of course, that’s only the beginning of the story, and CelesTrak still didn’t exist, so why did I take you on that part of the journey? The point was to show how your passion and curiosity—together with patience and perseverance—can take you to some incredible destinations. But what comes next will take us all to the next level. This is where my journey began to become our journey.
My first assignment after finishing my MS degree was to Sunnyvale AFS, CA to what we called the Blue Cube. It was where the USAF controlled all of its satellites. It wasn’t my first (or second) choice, but it ended up presenting some amazing opportunities.
One of the most fortuitous ones was that we were located right in Silicon Valley (Mountain View) as other computer users were trying to figure out how to connect their computers using things called modems and creating software for digital bulletin board systems (BBS). I got involved with this through one of the local computer clubs and started playing around with the software, trying to make it better.
As I was typing up the satellite data one night, I got this genius idea: What if when I typed up the satellite data, I put it on my BBS, so that other people could dial in and get it without having to type it up themselves? Surely there must be others that wanted to do that! And all they had to do was figure out what my home telephone number was, dial it up, and then download the data. As if!
Sometimes it pays to be a little clueless (or overly optimistic), but as it turned out, there were others back in 1985 that wanted to do just that.
Soon there was a growing universe of people sharing their interests, asking questions, and wanting to collaborate. At first they were just in Northern California, then across the US (long distance rates at the time didn’t help), and eventually even from other parts of the world. All calling my home computer.
Eventually, some started asking for more data, so I wrote NASA and asked for more. Then one day I got the response that I had reached NASA’s limit of 25 satellites and couldn’t get any more. So, I wrote them back and explained what I was doing and how that would be saving them paper, copying, and mailing costs. Soon they started sending me centimeter-thick packets once a week, just with TLEs on them!
Working together, we learned so much every day and did things many people couldn’t imagine. Our users all wanted to do more and the objective was to help everyone understand how to do these things easily. We started by providing standard, quality-checked data. We had already coined the terms “two-line element set” (it had been “two-card”) and TLE that are still widely used today. And I was constantly seeking ways to make it easier for us to continue to do more.
Most of that happened almost forty years ago—probably before many of you were born. But that was still only the beginning of our journey.
So, how does all of this relate to space operations?
Well, space operations was becoming a big part of who I was. I had gotten one of the first graduate degrees in it and had my first real taste of space operations while leading a team of 40+ people to launch and operate the Block I GPS satellites while at the Blue Cube.
And then I was heading off to the University of Texas at Austin to get my PhD so I could go back to AFIT to become a professor and teach space operations. While working on my PhD, I typed the entire “Spacetrack Report No. 3” into LaTeX (since PDFs didn’t even exist yet)—including the FORTRAN source code—which ultimately made SGP4 the de facto orbit propagator included in almost every piece of satellite tracking software.
And when I got to AFIT, I discovered the next big thing—access to the Internet.
Working with students who came from different parts of the space operations career field in the US, Canada, and Australia was a great way for me to stay on my toes and work to learn even more to help make the concepts of space situational awareness and astrodynamics relatable. And working as a space SME for various organizations helped, too. Seeing that need led to focusing CelesTrak on making orbitology more understandable and easier to apply. And soon (1994) we would be doing that via something called the World Wide Web.
And as CelesTrak continued to grow and help people get started in tracking satellites, many of those people ended up in some pretty amazing places and were doing some pretty amazing things. And they knew they could rely on CelesTrak to help them do many of those things, which has helped this journey continue.
As a result, I have gotten to work with amazing people all over the world. All the way from supporting the US team down at the South Pole since the early 1990s to working with ISRO on their record-breaking launch of 104 satellites on PSLV-C37 in February 2017.
CelesTrak was born from imagination and innovation and we continue to innovate to this day. Back when I was the colonel in charge of analysis for HQ AFSPC (2001-2003), very few people thought there was much chance that things could collide in orbit. The USAF screened its own satellites, but didn’t screen civil or commercial satellites, because that wasn’t their mission. And the assertion was that it would take a supercomputer and millions of dollars to produce results that ultimately wouldn’t be timely. But the US DOD relied on many of those commercial satellites to support our daily operations.
To me, that sounded like another challenge, so when I retired from the USAF in 2004, I set about to show that wasn’t true. By May 2004, we stood up SOCRATES, which screened all the payloads in Earth orbit against everything else up there, using standard data (TLEs) and COTS software/hardware (STK and a standard desktop computer) to produce results in an hour that looked out one week. And then we put those results on CelesTrak for everyone to see. That system still runs 3 times each day, almost 20 years later.
That changed the focus from conjunction analysis being too hard to do to then seeing that there were a LOT of close approaches happening all the time. Many satellite operators thought close approaches must be pretty rare since they had never been warned of one. Of course, they weren’t rare—but nobody was tracking them and issuing warnings.
Realizing there is a problem is the first step toward fixing it, and I was taught that problems are really opportunities to develop solutions to make things better.
In the beginning, we knew the TLE data wasn’t designed for conjunction analysis and the hardest things to track were the things that maneuvered. So, my first question was: What if we just asked the satellite operators who controlled those satellites to share their data? CelesTrak—working with Intelsat, SES, Inmarsat, and Eutelsat—started a prototype in 2008 that eventually became the Space Data Center (SDC). In less than a decade—through collaboration, transparency, and data sharing—we went from concept to FOC and set an example for the world of how to build international cooperation. By the time I retired (for the second time) in 2021, the SDC had 30-plus satellite operators, from about as many different countries, operating over 700 satellites.
And that brings us to today. And things are changing faster than ever and CelesTrak is continuing to lead the way.
CelesTrak has been moving the community forward since 2020 by implementing changes to the data formats we depend upon to support the growing population of satellites and debris in Earth orbit and making those formats readily available for software developers today while we continue to support the legacy formats.
We are using those formats to support easy visualization—graphs and, hopefully, 3D simulation again—of the data to make it more understandable—not just for us technical people, but for the policymakers and law makers, too. We need to ensure they can more easily grasp the concepts we work with to support their informed decision making.
And ultimately, everything we do is focused on the goals of maintaining safety of flight for space operations and protecting the Earth’s orbital environment. Achieving those goals is a team sport and it requires all of us working together to make it happen. Collaboration is key, as is transparency. And the best form of transparency is achieved through data sharing.
As in the beginning of our journey, data is the key. We know how to do most of the calculations, but we need to work together to make the data better suited to our needs and understand its limitations. CelesTrak works with satellite operators around the world—Starlink, OneWeb, Planet, Iridium, and many others—to do just that every day by fitting their data using SGP4 to produce supplemental GP data. This SupGP data allows us to positively know which satellite is which and provide the data just hours after launch. And CelesTrak provides SupGP data for over 5,400 of the 7,800 active satellites with GP data in Earth orbit or 69% of all active satellites. We need to grow that percentage together.
And we are working to make it easy to just use the data instead of having to put the data puzzle together first. Why should anyone still need to know what a TLE is, where to find it, and how to decode it, just to see what a satellite’s orbit looks like? We don’t do that when we use GPS on our smartphones to figure out where we are.
There’s still so much else to share of what we’re doing—we’ve actually barely scratched the surface of what CelesTrak is doing—but we will have to find another venue for that. But I hope this story—our story—excites you about the possibilities of what we can achieve through your curiosity, imagination, and dedication, and by working together. And it is all driven by our passion to learn and make things better. I hope we will find many ways to collaborate and work together to improve the safety, efficiency, and effectiveness of space operations going forward from today.
Mahalo nui loa—thank you very much—for letting me talk story with you and the honor of your attention. I look forward to any questions you have now and throughout this conference.