An environment to train and assess RL agents on the task of reaching a goal (x, y) position fast, smoothly and collision-free, navigating a simplified 2D aerospace in the presence of non-collaborative traffic.