This is a simple program, with the intent of using Machine Learning to predict possible collisions between satellites.
When running the program, you'll notice a processing bar in the terminal, this bar shows the rate which the satellites are being processed, and gives an estimate on how long the task would take:
When finished processing, plotly ensure to provide a 3d Graph, with the Earth plottet as a blue sphere, and the satellites as green dots. The line between the satellites show the possible collisions. Here is a image as reference:
If you wish to change the dataset and use you're own data to teach the ML and predict your own collisions follow these steps.
- Delete the file satellite_collision_model.joblib - This file provides the ML with the data it's already been taught
- Delete the file satellites.csv - This file holds the different satellites already processed and converted to a csv file.
- Navigate to Convert_data_to_csv.py
- In the script "Convert_data_to_csv.py" at line 10, a url is provided from which the data is recieved. Change this url to your disired API/own datasheet
It is important to note, that changing the csv_file name to something else, will require further changes to the main. Also this project is strictly for learning the use of Machine Learning in different scenarios. And certain flaws/bugs is to be expected.