Speaker
Description
UAV flight logs contain valuable information about flight trajectories, attitude changes, control inputs, sensor states, and system events. These records are essential for flight behavior studies, safety assessment, anomaly analysis, and data-driven UAV research. However, flight log analysis is often challenging because data formats vary across manufacturers, models, and flight-control systems. Raw log files may include inconsistent fields, irregular timestamps, missing records, and complex event structures, which can limit the reliability and reproducibility of scientific analysis. This presentation introduces a research software approach for UAV flight log extraction and scientific analysis. The proposed software support includes log parsing, data cleaning, field standardization, trajectory reconstruction, key-event identification, and visual analysis. Instead of focusing only on individual flight cases, the study emphasizes how research software can help transform heterogeneous raw flight records into structured and interpretable datasets for scientific investigation. Using representative UAV flight log samples, the presentation demonstrates the application of this approach in flight behavior interpretation, abnormal-event discovery, and comparative analysis across different cases. The work highlights the importance of research software engineering in improving transparency, repeatability, and efficiency in UAV data analysis, and provides practical insights for researchers working with complex drone flight data.