AKKF-based thermal UAV tracking targets edge deployment without giving up real-time speed

unimelb-nlp · hf · 2026-07-29

This paper proposes an edge-aware tracking pipeline for thermal infrared UAV swarm scenarios, centered on an Adaptive Kinematic Kalman Filter (AKKF).

The motivation is that tiny UAV tracking in thermal imagery needs both accuracy and edge-device efficiency. The authors combine AKKF with transient false-positive suppression and kinematics-driven predictive coasting to improve trajectory continuity under motion and sensor jitter. They evaluate the approach on the BSB benchmark and frame the work as a joint study of tracking quality and computational cost for real-time deployment.

Original post →

More from Research

Research channel →