Dataset for Ground-Vehicle Tracking of Fixed-Wing UAVs under Degraded Perception

This dataset supports the results reported in the manuscript “Ground-Vehicle Tracking for Dynamic Recovery of Gearless Fixed-Wing UAVs under Degraded Perception.” It contains numerical results from the Isaac Lab and CoppeliaSim evaluations of a perception-aware autonomous air–ground tracking framework. The Isaac Lab data include multi-method comparison and ablation experiments conducted under no, weak, and strong perception degradation. Five random seeds were evaluated for each configuration, with 128 parallel environments per seed. The reported data include success rate, mean maximum lateral error, 95th-percentile lateral error, root-mean-square error, and related tracking metrics. The CoppeliaSim data contain repeated closed-loop results obtained with ground-truth position, YOLO–LiDAR fused position, and fused position affected by latency and dropout. A README file describes the experimental conditions, file structure, variables, and units. The dataset is intended to support verification of the statistical results, figures, and tables presented in the associated article. Source code and third-party simulation assets are not included.