Interpretable Fuzzy Inference for UAV Target Tracking Using Bounding-Box Geometry

2026-08-04

AI Summary

This study addresses the challenges of perceptual uncertainty, high computational cost, and limited interpretability in continuous yaw angle estimation for ground targets by resource-constrained drones under visual guidance. To this end, the authors propose an interpretable fuzzy inference framework leveraging geometric features—center position, area, and aspect ratio—extracted from YOLO detection bounding boxes. The approach innovatively integrates Mamdani and Takagi–Sugeno fuzzy systems, constructing a compact rule base of only 27 rules derived from quantiles of the training data, thereby generating continuous yaw commands without explicit geometric modeling. Evaluated on 6,169 samples, the Takagi–Sugeno variant achieves a mean absolute error of 0.140°, 99.676% accuracy within ±1°, and 90.254% directional consistency, demonstrating high precision, strong interpretability, and suitability for real-time deployment.

Abstract

Vision-based guidance of unmanned aerial vehicles (UAVs) toward unmanned ground vehicles (UGVs) supports cooperative aerial--ground robotics, but reliable continuous yaw estimation from onboard vision remains challenging because of sensing uncertainty, limited computation, and the need for interpretable control. Existing deep-learning and geometric-reconstruction approaches often require large datasets, external localization, or complex modeling assumptions, reducing transparency and deployment suitability on resource-constrained platforms. We present an interpretable fuzzy-inference framework that generates continuous yaw commands from low-dimensional features extracted from YOLO boxes: target centroid location, area, and aspect ratio. No explicit geometric modeling is required. A Mamdani fuzzy system serves as an interpretable baseline using a shoulder--triangle--shoulder input partition. It is followed by a first-order Takagi--Sugeno model with three antecedent membership terms per input, whose parameters are derived from training-set quantiles, yielding a compact 27-rule structure. Evaluation uses 6{,}169 labeled samples from a VICON motion-capture environment. Across five randomized train--test splits, the Takagi--Sugeno model achieves a test-set mean absolute error of $0.140^\circ \pm 0.003^\circ$, a root mean squared error of $0.200^\circ \pm 0.008^\circ$, and a maximum absolute error of $1.254^\circ \pm 0.121^\circ$. Within-threshold accuracies are $99.676% \pm 0.270%$ for $\pm1^\circ$ and $100.000% \pm 0.000%$ for both $\pm3^\circ$ and $\pm5^\circ$. Directional consistency between image-plane horizontal displacement and predicted yaw sign reaches $90.254% \pm 0.612%$. These results show that the framework is transparent, data-efficient, computationally lightweight, and suitable for real-time vision-based UAV guidance toward mobile ground targets.

Problem

• UAV target tracking
• yaw estimation
• vision-based guidance
• interpretable control
• sensing uncertainty

Innovation

• interpretable fuzzy inference
• UAV target tracking
• bounding-box geometry
• Takagi–Sugeno model
• vision-based guidance