Structure-Aware Filtering for Real-Time Separation of Active-Sensing and Tracking Motions Aktif Algilama ve Takip Hareketlerinin Gerçek Zamanli Ayriştirilmasi için Yapi-Farkindalikli Filtreleme
34th Signal Processing and Communications Applications Conference, SIU 2026, İstanbul, Türkiye, 7 - 10 Temmuz 2026, (Tam Metin Bildiri)
- Yayın Türü: Bildiri / Tam Metin Bildiri
- Doi Numarası: 10.1109/siu71813.2026.11636672
- Basıldığı Şehir: İstanbul
- Basıldığı Ülke: Türkiye
- Anahtar Kelimeler: Active sensing, adaptive filtering, closed-loop control system, real-time signal processing, weakly electric fish
- Hacettepe Üniversitesi Adresli: Evet
Özet
Animals that rely on active sensing often generate high-frequency movements superimposed on slower task-related behaviors, making their functional role difficult to quantify in real time. We propose a structure-aware adaptive finite impulse response filter that separates the longitudinal motion of weakly electric fish into low-frequency refuge-tracking movements and high-frequency active-sensing oscillations. The filter is analytically derived and operates causally at the behavioral sampling rate. Simulation and ROS2-based real-time experiments show that the method selectively suppresses targeted frequency components while remaining stable under hardware constraints. Closed-loop experiments with freely swimming fish further demonstrate that manipulating the filtered high-frequency component systematically alters movement variability.