Variational Bayes methods for random matrix-based extended target tracking

Tuncer, Barkın
In this thesis, we investigate two major topics; Random matrix-based models using variational Bayes inference for extended target tracking and shape classification algorithms for tracking applications. In the scope of this thesis, we have derived two novel random matrix-based tracking algorithms. First, to represent the extent of dynamic objects as an ellipsoid with a time-varying orientation angle, and secondly, to estimate the extent of the object or a group of objects with more than one ellipsoid. In both of these solutions, we have used the variational Bayes technique to perform approximate inference, where the Kullback-Leibler divergence between the true and the approximate posterior is minimized by performing fixed-point iterations. The update equations are easy to implement, and the algorithms can be used in real-time tracking applications. We illustrated the performance of the methods in simulations and experiments with real data. In extended target tracking, once the shape estimate of an object is formed, it can naturally be utilized by high-level tasks such as classification of the object type. Therefore, we present a naively deep neural network, which consists of one input, two hidden and one output layers, to classify dynamic objects regarding their shape estimates. In this manner, the proposed method shows superior performance in comparison to a Bayesian classifier for simulation experiments.


New methods for decentralised sensor fusion and extended target tracking models
Köksal, Hilal; Özkan, Emre; Department of Electrical and Electronics Engineering (2022-12-16)
The focus of this thesis is two-fold; the first study investigates the use of student-t distribution for the decentralised multi-sensor fusion problem. Multi-sensor fusion can suffer from several artefacts such as low channel capacity, delays in the communication channels, the correlation in the acquired information and sensor biases. When combined with these artefacts, the errors in the local sensors’ estimates can result in conflicting information at the fusion center. Such conflicts may later be resolved...
Variational smoothing for extended target tracking with random matrices
Kartal, Savaş Erdem; Orguner, Umut; Department of Electrical and Electronics Engineering (2022-4-05)
In this thesis, two Bayesian smoothers are proposed for random matrix based extended target tracking (ETT). The proposed smoothers are based on the variational Bayes techniques and they are derived for an extended target model without and with orientation. The random matrix models of Feldman et al. and Tuncer and Özkan are used as the extended target models without and with orientation, respectively. The performance of both smoothers is evaluated using simulation results on two different scenarios. It is se...
Evaluating an Immersive Virtual Learning Environment for Learning How to Design in Human-Scale
Sonmez, Orkun; Sorguç, Arzu (2018-01-01)
This paper presents a part of a thesis research conducted at METU. It proposes a method for evaluating the effects of an immersive virtual learning environment (IVLE) which is integrated in an architectural design/learning activity. Proposed IVLE application and design/learning activity were designed through a synthesis on constructive learning, problem-based learning, immersive technologies, and intended learning outcomes (ILOs) in learning how to design in human-scale. Immersive experience of bodily inter...
Second-order scalar-tensor field theories
Şahin, Ertan Sinan; Sarıoğlu, Bahtiyar Özgür; Department of Physics (2017)
We review Horndeski’s scalar-tensor theory in this thesis. Partial differential equations that are satisfied by the Lagrangian limit its most general form. Demanding second-order field equations both for the metric and the scalar field, and choosing a four-dimensional spacetime also put restrictions on the most general form of the Lagrangian. Besides, by using similar techniques, in a four-dimensional spacetime, we find the most general form of the second-order Euler-Lagrange equations that are obtained fro...
Framing Effects On New Generation Financial Statements.
Gönül, Mustafa Sinan; Muğan, Fatma Naciye Can; Akman, Nazlı(2013-12-31)
We will investigate the perceptions about the new format of the financial statements and evaluation of it by both expert and novice groups will also be gathered. This will provide a glimpse on how easy/difficult the adoption of this new standard will be when it will be forced into implementation. To investigate the effects of the aforementioned factors on the perceptions and judgments of the financial statement users, an experiment having a 2x2x4 factorial design will be employed. The independent variables ...
Citation Formats
B. Tuncer, “Variational Bayes methods for random matrix-based extended target tracking,” M.S. - Master of Science, Middle East Technical University, 2022.