Quadcopter Flight Trajectory Control Dataset

This dataset comprises real-time dynamic response data consisting of 10-dimensional state variables and the corresponding 3-dimensional control inputs during quadcopter operation. 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. These models can be applied to trajectory evolution analysis, flight control law mining, and digital twin simulations.

本数据集包含四旋翼无人机（quadcopter）运行过程中的实时动态响应数据，该数据由10维状态变量与对应的3维控制输入构成。其中状态变量由动作捕捉系统（motion capture system）实时采集，涵盖无人机的位置、速度与姿态四元数；控制输入则由板载计算机在线计算生成，对应三轴加速度控制指令。本数据集面向数据驱动与智能轨迹控制领域的研究，可为深度神经网络、强化学习控制策略的开发以及系统辨识任务提供高质量的训练与测试样本，还可支撑复杂三维环境下智能控制律与规划算法的设计与验证工作。此外，飞行状态与控制输入间的时序映射关系可用于构建高精度的四旋翼无人机动力学与运动学模型，此类模型可应用于轨迹演化分析、飞行控制律挖掘以及数字孪生（digital twin）仿真等场景。