Development of an intelligent dynamic stabilization system for a drone dedicated to collecting beached marine waste

CONTEXT Autonomous drones represent a promising solution for collecting marine waste stranded on beaches and coastal areas that are difficult to access. The dynamic stabilization of the drone during collection is a major challenge due to uneven terrain, external disturbances (wind, waves) and load variations. INTERNSHIP OBJECTIVES • Develop an intelligent stabilization system based on AI • Compare classical approaches (PID) and intelligent approaches (RL, neural networks) • Validate in simulation under realistic conditions TASKS REQUIRED • Bibliographic study (PID, fuzzy logic, RL, neural networks) • Design of the stabilization system architecture • Development of the stabilization model (RL/Deep RL or ANN) • Simulation under ROS/Gazebo with realistic disturbances • Comparative analysis of classical PID vs AI approaches DELIVERABLES • Intelligent stabilization model • ROS/Gazebo simulation environment • Detailed comparative analysis • Scientific article