Quadcopter Racing Simulation Dataset

该数据集包含了为四旋翼无人机竞速设计的神经网络控制器所需的训练和评估数据，这些数据在3英寸和5英寸的无人机上进行了测试，并且应用了不同级别的域随机化。此外，数据集还包括了在不同域随机化水平（0%、10%、20%、30%）的模拟结果，以及真实世界测试的性能指标。该数据集涵盖了100百万时间步长，跨越了9个不同的环境，其任务是进行四旋翼控制的神经网络训练。

This dataset provides training and evaluation data for neural network controllers tailored for quadrotor drone racing. The data were tested on 3-inch and 5-inch drones with varying levels of domain randomization. Furthermore, the dataset includes simulation results across four domain randomization levels (0%, 10%, 20%, 30%) and performance metrics from real-world tests. Spanning 100 million time steps across 9 distinct environments, this dataset is designed for training neural networks for quadrotor control tasks.