A Survey on Reinforcement Learning Methods for UAV Systems

What is it about?

In recent years, Unmanned Aerial Vehicles (UAVs) have attracted a lot of attention due to their flexibility and mobility. However, due to the increasingly complex environments faced by UAVs and the rising demands on UAV systems, traditional UAV control methods can no longer efficiently control the UAV under multi-constraint situations. Reinforcement Learning (RL), as an emerging robot control technology, is well suited to the needs of UAV systems in terms of its ability to interact with and learn from the environment. Therefore, RL-based UAV systems are gradually becoming a new trend in research. Nonetheless, as a new research field, it faces some challenges. To fully grasp the landscape of RL-based UAV systems, it is paramount to provide a comprehensive overview and analysis of the existing specific RL methods applied to UAV systems. In this survey, we first provide a comprehensive overview and summary of the application of RL in different UAV scenarios based on the classification of RL methods. After that, based on the existing relevant literature, we conduct a systematic analysis of the challenges and recent advancements when applying RL to UAV systems. Finally, we discuss the potential research directions for RL-based UAV systems.

Featured Image

Photo by George Kroeker on Unsplash

Photo by George Kroeker on Unsplash

Why is it important?

Unlike previous surveys that mainly focus on multi-UAV systems, our work also examines single-UAV studies, offering a more comprehensive view of RL applications in UAVs. We categorize existing research into value-based, policy-based, and Actor-Critic (AC) methods and discuss key challenges such as high-dimensional spaces, limited observations, and dynamic environments. Finally, we outline promising future directions including cooperative control, sim-to-real transfer, and the integration of LLMs.

Perspectives