Sıralı Monte Carlo kullanarak elips olmayan genişletilmiş nesnelerin takibi.

---

Tezin Türü: Yüksek Lisans

Tezin Yürütüldüğü Kurum: Orta Doğu Teknik Üniversitesi, Mühendislik Fakültesi, Elektrik ve Elektronik Mühendisliği Bölümü, Türkiye

Tezin Onay Tarihi: 2018

Tezin Dili: İngilizce

Öğrenci: Süleyman Fatih Kara

Danışman: EMRE ÖZKAN

Özet:

The problem of extended target tracking is considered in which the target extent is represented with multiple ellipses. The resulting inference problem, which is considered in the sequential Monte Carlo (SMC) framework, includes association of the measurements between sub-objects. We make use of different particle filtering approaches to solve the aforementioned association problem under the assumption of known extent. When the extent is unknown, parameters of the multiple ellipses should also be estimated. For this purpose, a particle filter based method is derived for joint estimation of target's kinematic and extent states. The proposed method uses variational Bayes technique to obtain an approximate conditional analytical expression, which enables the use of Rao-Blackwellization (a.k.a. marginalization) idea in particle filtering.