20.500.12556/RUL-183221 EKF-based target estimation for the APN missile guidance using a YOLO-driven seeker Ocenjevanje cilja z razširjenim Kalmanovim filtrom za vodenje rakete APN z uporabo iskalnika YOLO Sodobni zakoni vodenja takticnih raket zahtevajo natan ˇ cno ˇ sledenje cilju in ocenjevanje njegovega stanja za ucinkovito ˇ prestrezanje manevrirajocih ciljev. Proporcionalna navigacija ˇ je eden najpogosteje uporabljenih zakonov vodenja, ki temelji na meritvah hitrosti spremembe smerne crte za generiranje ˇ bocnih pospeškov. Rakete obi ˇ cajno uporabljajo elektroopti ˇ cne ˇ ali infrardece iskalnike. V naši raziskavi uporabljamo dvoosni ˇ kardaniziran kamerni iskalnik z laserskim merilnikom razdalje, opremljen z algoritmom YOLO za zaznavanje, sledenje in prepoznavanje ciljev v realnem casu. Iskalnik omogo ˇ caˇ zajem slike, zaznavanje cilja, sledenje ter merjenje hitrosti spremembe smerne crte glede na inercialni koordinatni sis- ˇ tem. Poleg tega uporabljamo razširjeni Kalmanov filter za filtriranje meritev in oceno tridimenzionalnega gibanja cilja ter bocnih pospeškov. Te ocene nato uporabimo pri razširjeni ˇ proporcionalni navigaciji, ki v primerjavi s standardno zmanjša potrebne bocne pospeške rakete. Rezultati raziskave so potrjeni ˇ z nelinearnim modelom rakete s šestimi stopnjami prostosti in kažejo ucinkovitost povezovanja sledenja cilju na osnovi ˇ globokega ucenja s sodobnimi zakoni vodenja. Modern tactical missile guidance laws require an accurate target tracking and state estimation to effectively intercept maneuvering targets. The Proportional Navigation (PN) is one of the most widely used guidance laws for tactical missiles, relying on line-of-sight (LOS) rate measurements to generate lateral acceleration commands. Missiles typically employ electro-optical or infrared seekers. Our study utilizes a two-axis gimbaled camera seeker which employs a laser range finder, equipped with the YOLO algorithm for a real-time target detection tracking and recognition. The seeker provides the image acquisition, target detection, tracking, and LOS rate measurement relative to the inertial frame. Additionally, an Extended Kalman Filter (EKF) is used to filter the LOS rate measurements and to estimate the target three-dimensional motion, as well as lateral accelerations. The estimates are then used to implement the Augmented Proportional Navigation (APN) which reduces the required missile lateral acceleration compared to the standard PN guidance. The study findings are validated using a six-degrees-of-freedom (6-DOF) nonlinear missile model, demonstrating its effectiveness when integrating the deep-learning-based target tracking with the modern guidance laws. EKF missile guidance YOLO 6DOF APN seeker true false true Slovenski jezik Angleški jezik Članek v reviji 2026-06-08 14:45:39 2026-06-08 14:45:47 2026-06-09 14:12:53 0000-00-00 00:00:00 2026 0 0 Str. 44-55 št. 1-2 Letn. 93 2026 0000-00-00 NiDoloceno NiDoloceno NiDoloceno 0000-00-00 0000-00-00 0000-00-00 623.465.2:681.518.5 0013-5852 280851203 RAZ_Hodzic_Mirza_2026.pdf RAZ_Hodzic_Mirza_2026.pdf 1 536577BD0AC4426619453D0037E78021 69588f1350286cfa20a6a5b564cad902e60c2375a8314deabe4d39f879d95f25 f741e6d8-6337-11f1-9b0d-0050569b8976 https://repozitorij.uni-lj.si/Dokument.php?lang=slv&id=235109 https://ev.fe.uni-lj.si/1-2-2026/Hodzic.pdf 0 5817550a-63fb-11f1-9b0d-0050569b8976 https://repozitorij.uni-lj.si/Dokument.php?lang=slv&id=235204 Fakulteta za elektrotehniko Elektrotehniški vestnik 0 0 0