Unmanned aerial vehicles (UAVs), or drones, are increasingly deployed for critical missions such as environmental monitoring, infrastructure inspection, and disaster response. Assessing the reliability of these missions is essential for operational planning, yet conventional approaches often fail when input data are incomplete or epistemically uncertain. We present a novel framework for mission availability analysis that integrates fuzzy decision tree (FDT) induction with binary decision diagram (BDD) construction. The method interprets a drone mission as a reliability system, where checkpoints act as components and mission success is modeled by a structure function. Expert evaluations expressed as confidence degrees are used to induce an FDT, which is subsequently defuzzified and transformed into a canonical BDD. This representation enables efficient computation of mission availability and sensitivity measures using established BDD algorithms. We validate the approach on a real-world case study of a forest fire monitoring mission comprising eight checkpoints and demonstrate high predictive accuracy (94%) despite incomplete training data. The proposed method provides a transparent, reproducible pipeline for translating uncertain, expert-driven data into quantitative reliability metrics, offering practical insights for mission planning under uncertainty.
Zaitseva et al. (Sat,) studied this question.