Controlled Space Debris with Machine Learning

Over the years, the Space junk has become a major problem for upcoming space missions. Its importance is accentuated when studying the speeds with which scrap metal travels, since it can cause unwanted collisions with equipment that is being launched recently.

To mathematically calculate the progression of these shocks, we use the artificial intelligence, known as Machine Learning. This helps to calculate possible shocks using a machine learning considering an input of information into your system.

What is space junk?

Space junk is any object or scrap that orbits and travels through space without a defined utility or function. It is composed of old satellites, abandoned rockets, the remains of a space station or external components such as dust or paint.

This scrap comes from explosions or the detachment of the different stages of a rocket. It also consists of technological elements designed by man to be launched into outer space.

Although the various ships and exploration equipment have helped to study the exterior, they have also been a constant source of waste.

Nowadays it is estimated that There are more than 7200 tons of useless fragments orbiting in space, constituting a major problem for future space missions.

Damage caused by space debris

Space began to be contaminated approximately 60 years ago, as a result of space exploration to study the unknown. Although many missions have been achieved, a large amount of waste which could currently be causing some damage. Among the most important are:

1. Accumulated garbage is dangerous because it is against growth. As a result, it increases in size and quantity, which can be detrimental to the planning and orbit of any equipment launched into space.
2. Although many debris is tiny, the fast speed they acquire in space makes them projectiles Very dangerous. These can affect the trajectory of any mission, even causing a malfunction.
3. This garbage endangers the lives of astronauts and aerospace professionals. Especially when there is a risk of collision with some debris navigating nowhere.
4. Some satellites have been threatened by debris surrounding them. These can produce irreversible damage to equipment, in the long run, it would only generate more space junk.
5. For the terrestrial population, the risk of collision is lower, although it is still a possibility. Over the years, fragments of rockets and space plants have been found in various parts of the world.
6. With damaged satellites, telecommunications, global positioning and general defense systems can be affected. Even climate services would be affected by this.
7. Finally, a problem that encompasses all of the above is the Kettler syndrome. This means that all debris collisions create new garbage, which is thrown back into space to collide again in the future. Therefore, it is a chain reaction that can cause no more satellites or space flights to be sent outside the Earth.

Although they have not caused major damage, the dangerousness of these ingredients is still alarming.

Description of the problem and evaluation of the solution

As is already known, there are many problems deriving from space debris travel in our orbit. For this reason, multiple options have been put on the table to solve problems that may affect us in the future. However, there are several aspects to consider, starting with the equipment and human resources that are needed.

For this reason, some feasible methods may include the use of artificial technology to assess the impact or potential collisions of useless traveling bodies.

The problem

When objects are propagated through space, the risk of collision between space and other matter traveling on that trajectory increases. As a result, account must be taken of the possibility that this junk will collide with a major satellite when it enters its orbit.

• As a result, the equipment may receive irreparable damage, which would entail a technological and economic loss for the nations that launched the launch.
• On the other hand, this also means a significant loss of information, which may be important for the advancement of new space research.
• Therefore, with the use of artificial intelligence, Machine Learning, we seek to reduce the level of uncertainty regarding future collisions. In other words, when a possible crash is established, with this program, a model can be built that uses all the information obtained to predict the final risk.
• The collection of all the information needed for this program is set out in a conjunction message (CDM). Each of these messages contains important information about the possible situation, such as the affected satellite, the nature of the space debris and the time at which the collision may occur.

Evaluation of the solution

With the Machine Learning technology it seeks to calculate the possible collisions between two specific objects. For this, the data stored in the CDM (Conjunction Data Message) is used. The last of these messages contains all the data of the next commission, which will be analyzed by the Space junk office.

This office is responsible for alerting control teams to evaluate maneuvers to avoid collisions. These will be put into operation 1 day before the crash.

The solution is measured according to the mean square error (MSE) and F2.

• The squared error indicates the reliability of the risk prediction. Therefore, it is a relationship between calculated risk and actual risk. However, this error has several meanings and the margin of error may vary depending on the scale of the risk.
• As for F2, it is responsible for evaluating and ruling out unlikely collisions. By using F2 with the average error, the solutions for the really high risks are determined.

As a result, the techniques used by offices in charge of space debris seek Solving a growing problem. With the next space missions, the level of scrap in space could increase, which would have long-term consequences. For this reason, more studies are being carried out every day to achieve, by various means, the solution to this great problem.