Researchers have developed an interpretable fuzzy inference framework for guiding unmanned aerial vehicles (UAVs) toward ground targets. This system extracts low-dimensional features from YOLO bounding boxes, such as centroid location, area, and aspect ratio, to generate continuous yaw commands without requiring explicit geometric modeling or large datasets. A Takagi-Sugeno fuzzy model with 27 rules achieved a mean absolute error of 0.140 degrees in a VICON motion-capture environment, demonstrating its suitability for real-time, resource-constrained applications. AI
IMPACT This interpretable fuzzy inference system offers a lightweight and data-efficient approach for real-time UAV guidance, potentially improving autonomous navigation in complex environments.
RANK_REASON Academic paper detailing a new method for UAV target tracking. [lever_c_demoted from research: ic=1 ai=1.0]
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
