• Tracking

The discussed work pertains to the research of a method for tracking space debris and satellites using Artificial Intelligence Methods. By utilizing two neural networks, a smoothing of the trajectory can be achieved, and it can be predicted in such a way as to be used in radar systems. With regards to the first architecture (the one that performs smoothing on the target trajectory), it is observed that as the signal to noise ratio increases, or in other words, as the body being considered becomes larger, the performance of the network deteriorates. However, the aforementioned method can be improved with further training and is highly adaptable to different types of signal. Concerning the neural network for predictions, it is evident that it is notably able to predict future radar detection, and through filtering methods, the system is useful for potential uses of collision avoidance and radar targeting.