CS 171: Final Project Write-up by krzysztof uscinski

System requirements.

Back to Top

What are your motivating questions and why did you choose them?

My main goal of my visualization was to find out if there are any patter differences between space debris and space satellites currently orbiting earth. My prediction is that even space debris will have some orbital pattern, generated from Earth's gravitational pull. Active space satellites will have a more defined orbital pattern, where then nicely circle earth, the advantage being that they are still powered by jets and some sort of steering, so they stay in orbit where they are supposed to.

I chose this question because space seems very chaotic, and it seems that there were a lot of satellites launched, and a lot of satellite crashes. I am curious to see what the various patters look like for active satellites, versus inactive satellites and space debris.

Back to Top

Describe your data in detail: where did it come from, how did you acquire it, what does it mean, etc.

The data used for my visualization is based on Two-line element sets (TLEs).

As defined by Wikipedia, Two-line elements are:
A Two-line element set (TLE) is a set of Orbital elements that describe the orbit of an earth satellite. A computer program called a model can use the TLE to compute the precise position of a satellite at a particular time. The TLE is in a format specified by NORAD and used by NORAD and NASA. The TLE can be used directly by the SGP4 model (or one of the SGP8, SDP4, SDP8 models). Orbital elements are determined for many thousands of space objects by NORAD and are freely distributed on the Internet in the form of TLEs.[1] A TLE consists of a title line followed by two lines of formatted text.

In other words, TLEs are a set of data points used to PREDICT where specific satellites are in space. To make the data more useful, it helps to convert TLEs into Longitute, Latitude and Altitude, so each point can be plotted with relation to Earth.

Converting TLEs to Lon, Lat and Alt data takes a lot of math. The math is faily extensive and it would take a lot of time for me to come up with a model for the conversions. I decided to find a source of converted TLEs. This place was www.agi.com .

I will explain in detail the process of mining TLE data sets below.

The above image shows the PREFFERED method of using TLEs. The first step would be aquiring the TLEs. There are some public services on the web that offer TLE data sets. wwwcelestrak.com is one of these sites. The next step would be to take the TLEs and run them though the complicated math to derive the langitude, latitude and altitude for each data point. The converted TELs would then be saved in an XML file, specifically .KML which is a Google Earth format. The .KML files have a specific format that Google Earth can read and display in the Google Earth program. I decided to use this same format because I could check my visualization against Google Earth. Once the TLE data sets are converted into .KML XML format, they can be used by my visualization. This is the hard way of displaying TLE data sets. Next, I shall talk about the easier way, the way I chose, to display the TLE data sets in my visualization.

The above image shows the EASIER way of displaying TLEs. This is the way I chose for my visualization. In a nut shell, this method simply uses someone else's converted TLEs. This source for me was www.agi.com . AGI provides commercial off-the-shelf software to national security and space professionals for integrated analysis of land, sea, air, and space assets. The technology can be applied to a variety of solution areas. AGI provides real-time TLE conversion in .KML format. These data sets can be found at http://adn.agi.com/SatelliteDatabase/KmlNetworkLink.aspx . This link allows you to download SatelliteDatabase.kmz with is a Google Earth native format. The .KMZ file is really a ZIP file. The next step after downloading the .KMZ file is to rename it to .ZIP . Extracting all the files from the .ZIP file is the next step. The extraction will generate about five files, one of them being doc.kml . Doc.kml is the actual XML formatted data set. Doc.kml is what gets imported by my visualization.

Weather method 1 or 2 are followed to convert the TLE data sets into longitude, latitude and altitude formats, the resulting DOC.KML file will allow us to view each satellite or space debris relative to its' position above Earth.

Back to Top

Describe your design and why you chose the features you did.

My visualization starts out with the information screen. It simply states what your options are, and how to navigate though the visualization. It also tells you what mouse buttons to use for particular tasks.

The second scree, the 2D screen, displays the satellite data points in 2D mode, which is simply mapping all the satellites in 2D on a map of the world. In this mode, all points are encoded with circle sizes determined by the satellite's altitude. The higher the satellite, the smaller the circle and the lower the altitude of the satellite, the larger the circle will be. The color of the circles is transparrent, so that you can see overlapping data points. This was done becasue when all categories are selected, all the data points blend together too much. the 2D screen and the 3D screen have the categories buttons. These buttons allow the user to select the data sets. The data sets available are Active Satellites, Inactive Satellites, Space Debris and Rocket Bodies. Each category is encoded with a color. The 2D and 3D screen also have an Owner filter, which allows you to display data points based on countries that launched the satellites. The Owner filter also has a visual sniffer that shows a number of satellites in specific categories. The 2D mode has a legend that explains the circle size encoding. Finally, both 2D and 3D modes have Double Sliders to help filter out the Year the satellites were launched, the Longitude, Latitude and the Altitude of the satellites. This allows to tailor searches. The Year double-slider also has another visual sniffer. This sniffer shows the frequency of satellites for specific years along the Year slider. The user can then slide along the Year slider, stopping on specific years that may be of intrest. One thing to point out about the double-sliders is the fact that they can specify the lower-limit and an upper-limit for the filters. Hovering over the data points displays more information. Clicking on data points highlights the data point and allows the user to view the highlighted data point in the 3D mode.

The 3D screen is exactly the same as the 2D screen except it lacks the altitude legend and the data points are plotted in 3D. The user can then rotate Earth by left-clicking and dragging the screen, and zoom in and out by right-clicking and dragging the screen. Highlighted points from the 2D screen show up as satellite name labels that rotate with the Earth. The user can easily hide the filters in both 2D screen and 2D screen by clicking on the Filters button.

Back to Top

How do you use your application (mouse and keyboard functions, input/output, etc)?

Back to Top

What are the answers to your questions? What other interesting insights about your data did you gain from your visualization?

The active satellites definately have a clean pattern of satellites orbiting Earth. There is a nice ring above the equator, with some lower altitude orbiting satellites.

The space debris, on the other hand, has no pattern in general. It's just a clump of specs on the screen... until...

When the filters are applied to extract the right data points, one can see that even the space debris has an orbiting pattern. In the images above, we can see how the debris orbits Earth due to Earth's gravitational pull.... interesting.

Back to Top

What extensions and improvements can you suggest?

The one thing I wish I would have done is my model to convert the TLEs data sets into Longitude, Latitude and Altitude. I acutally started working on a converter, but didn't have enough time to finish. This would be my only recomendation for improvement.

Back to Top

What did you most enjoy about working on this project? What was the most challenging aspect? What was the most frustrating? What would you do differently next time?

The most frustrating issue was using Processing, because it still has some bugs in it.

Back to Top

The Visualization...

here...

Back to Top