Recursive identification algorithms for jump Markov linear systems and elliptical extended target tracking

Date

2025-1-09

Author

Çetinkaya, Mehmet

Metadata

Show full item record

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

Item Usage Stats

85
views

0
downloads

Cite This

This thesis addresses two challenges in real-time applications: the online identification of jump Markov linear systems (JMLSs) and the recursive estimation of elliptical target extents. For JMLSs, we develop a computationally efficient recursive expectation-maximization (EM) algorithm that leverages the conditionally linear structure of these systems, enabling higher-dimensional model identification without particle filters. State inference is performed using a multi-model filter, with simulations validating the ability of the method to identify transfer functions, noise parameters, and transition probability matrices. Compared to state-of-the-art batch EM methods, our approach offers significantly reduced computational cost, making it suitable for real-time applications. For elliptical extended target tracking, we propose a unified framework that estimates target extent lengths and orientation. This includes a recursive EM algorithm, a log-likelihood linearization-based recursive algorithm, and a novel iterated extended Kalman filter that processes measurements in batches while maintaining order invariance. Simulations and real-world experiments demonstrate superior accuracy and efficiency compared to existing techniques.

Subject Keywords

Jump Markov linear systems, Recursive EM algorithm, Online identification, Elliptical target tracking, Iterated extended Kalman filter

URI

https://hdl.handle.net/11511/113460

Collections

Graduate School of Natural and Applied Sciences, Thesis