Python Code and Data Supporting Chapter 5 of the PhD Thesis: Symmetry-informed Offline Learning for Flight Control

The dataset contains relevant Python code and supporting data in relation to CHAPTER5 of the thesis Safe and Sample Efficient Reinforcement Learning for Flight Control by Yifei Li (2026). The thesis aims to develop reliable and efficient online and offline RL algorithms for data-driven tracking control applications. This chapter concerns exploiting symmetric data augmentation technique to improve sample efficiency of deep deterministic policy gradient (DDPG), and the application on lateral flight control problem. The code provides DDPG algorithms and vehicle dynamics simulation framework, and visualization of training process. The PyTorch package serves as the main instrument for simulation. For more information, we refer to the respective thesis chapter.