Quadcopter Flight Trajectory Control Dataset

The state variables, captured in real-time by a motion capture system, include the drone's position, velocity, and attitude quaternions. The control inputs, computed online by the onboard computer, represent acceleration commands across three axes.The dataset is designed for research in data-driven and intelligent trajectory control.

It provides high-quality training and testing samples for developing deep neural networks, reinforcement learning control strategies, or system identification. It further supports the design and validation of intelligent control laws and planning algorithms in complex 3D environments. Additionally, the temporal mapping between flight states and control inputs enables the construction of precise quadcopter dynamic and kinematic models.