Data used for the training and evaluation process of the ML model presented in "Machine Learning-based resolution of strategic conflicts in U-space airspaces"

This repository contains the datasets used to evaluate and validate a machine learning–based conflict detection and resolution framework for UAV traffic operating in a realistic PX4-based simulation environment. The data include both real-world-inspired operational scenarios and synthetic grids generated exclusively for model training. Together, these files provide the full context required to reproduce the experimental setup, evaluate model performance, and compare scenarios across different environments. table_cells.csv and times.csv contain information about the real environment, which are the scenarios simulated by the PX4 realistic simulator. The first contains the cells that have been created in this area and the second contains the cells that have been occupied, and at what time they are occupied. This are the environments where the ML tool is evaluated. real_grids_21x28x194_n5.npy are the grids that have been created only for the ML training process. ois.json are the operational intents of the flights that have been rejected by the Flight Authorisation Service (FAS). This is the data that has been used for the evaluation of the model.