National Technical University of Athens (NTUA)
Αναδρομική εκτίμηση παραμέτρων σε γραμμικά και μη γραμμικά μοντέλα εφαρμογές στην πρόβλεψη ανάπτυξης καρκινικών όγκων και την παρακολούθηση της θαλάσσιας κίνησης
Abstract
dc:descriptionIn this thesis, a variety of recursive estimation techniques for parameter and state estimation on linear and non-linear models were applied. These techniques make use of the information contained in the measurements in order to estimate the unknown parameters. Two seperate applications were studied. The first one deals with tumor growth modeling and prediction based on the tumor's volume measurements, while the second one deals with maritime surveillance, vessel monitoring and abnormal behavior detection. As regards tumor growth modeling, a methodology for the estimation of the unknown parameters of the non-linear Gompertz growth model is presented. This model is commonly used in the literature to describe the growth of macroscopic tumor measurements. According to the Gompertz model, the growth of a solid tumor is characterized by two main features, the growth rate and the carrying capacity. In the literature, when fitting experimental tumor data to the Gompertz equation, the carrying capacity of a tumor is considered as a species--specific parameter, and thus has a constant value. However, the theoretical assumption that any given tumor will approach the same asymptotic value has not been proven experimentally. On the other hand, the heterogeneity of tumor growth has been recognized to have clinical implications in the planning of effective treatment schedules. The methodology presented in this thesis considers both the growth rate and the carrying capacity to be subject--specific parameters and uses estimation techniques based on the principle of Maximum Likelihood in order to estimate their values. It has been observed that this approach can provide more accurate estimates of the unknown parameters for the synthetic data and more accurate predictions of the future growth for the experimental data. On the topic of maritime surveillance, this thesis focuses on estimating the route of a vessel and detecting abnormal behaviors based on AIS data. A software application was created in order to detect these behaviors and notify the user whether or not further investigation is required. Linear Filtering using AR models is proposed in order to create trajectory forecasts. Furthermore, various alert criteria are set expressing different abnormalities. AIS data collected from Saronikos Bay were used and the results showed that the program could successfully monitor the bay in real time and report abnormalities in vessel behavior.
Degree
thesis:*- Grantor dc:publisher
- National Technical University of Athens (NTUA)
- Year dc:date
- 2019
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Patmanidis, Spyridon
- Πατμανίδης, Σπυρίδων
Subjects
dc:subject × 28- Αναδρομική εκτίμηση παραμέτρων
- Αναδρομική εκτίμηση καταστάσεων
- Γραμμικά συστήματα
- Μη γραμμικά συστήματα
- Μέγιστη πιθανοφάνεια
- Ελάχιστα τετράγωνα
- Φίλτρο Καλμαν
- Φίλτρα σωματιδίων
- Βιολογικά συστήματα
- Μοντελοποίηση καρκινικών όγκων
- Παρακολούθηση θαλάσσιας κίνησης
- Recursive parameter estimation
- Recursive state estimation
- Linear systems
- Nonlinear systems
- Maximum likelihood estimation
- Least squares estimation
- Kalman filters
- Particle filters
- Biological systems
- Tumor growth modeling
- Maritime surveillance
- Επιστήμες Μηχανικού και Τεχνολογία
- Επιστήμη Ηλεκτρολόγου Μηχανικού, Ηλεκτρονικού Μηχανικού, Μηχανικού Η/Υ
- Βιοϊατρική Μηχανική
- Engineering and Technology
- Electrical Engineering, Electronic Engineering, Information Engineering
- Medical Engineering
Rights
dc:rights- Statement dc:rights
-
- BY_NC
- Language dc:language
- gre
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/45436