Fitting dynamical models to data

Uitgebreide vaknaam	Fitting dynamical models to data
Leerdoelen	This course aims at providing background and mathematical tools for fitting parameters in a dynamical model to available data. It complements courses on the modeling of (complex) systems. This course is also suitable for anyone who will be involved in the analysis, optimization and control design of dynamical systems, including, electro-mechanical systems, chemical processes, operations, biomedical systems and biological systems. At the end of the course, students will be able: LO1: to construct an ARMAX or IIR or discrete-time state-space model from input-output signals; LO2: to estimate/to identify the unknown parameters in a dynamical model, which fit with the available data, using various computational tools; LO3) to refine/to adapt the parameters recursively based on newly acquired data.
Omschrijving	Dynamical modeling has played an important role for the analysis, optimization and control design of systems, including, electro-mechanical systems, chemical processes, operations, biomedical systems and biological systems. The models are generally constructed based on physical laws or phenomenological behavior. These models contain parameters which need to be identified in order to capture the essential systems dynamical behavior. In this course, the students will learn methodologies for estimating the parameters in the models using the available data. The topics that are covered in the class include: - Introduction to difference equations - Auto-correlation and cross-correlation functions - Wiener filter - Least-mean square filter - Recursive Kalman filter and recursive least-square filter - Maximum-likelihood filter - Particle filter and extended Kalman filter - Bayesian filter - State-space realization theory
Uren per week
Onderwijsvorm	Hoorcollege (LC), Practisch werk (PRC) (The practicals are computer practicals.)
Toetsvorm	Opdracht (AST), Schriftelijk tentamen (WE) (individual asignment, final exam with open questions)
Vaksoort	master
Coördinator	prof. dr. ir. B. Jayawardhana
Docent(en)	prof. dr. ir. B. Jayawardhana
Verplichte literatuur	Titel Auteur ISBN Prijs Lecture Notes on Fitting Dynamical Models to Data
Entreevoorwaarden	The course unit assumes prior knowledge acquired from Signals and Systems for IEM/BMT, Mechatronics and Numerical Methods. This course prepares student for the IEM IEM Research Master Thesis project.
Opmerkingen	There will be a group assignment. Every group has to decide on their own, on the systems that will be identified. Each group has to collect the data themselves. Each group will present their system, the data and the methods in the class room from Week 3 onwards. Take-home test exam: the sheet will be distributed at the last lecture. Final grade = max{take-home test grade, 0.60 x individual assignment + 0.40 take-home test grade}. This course was registered last year with course code TBAFPE-11
Opgenomen in	Opleiding Jaar Periode Type MSc Computing Science: Data Science and Systems Complexity  (Guided choice course units) - semester I b keuze MSc Courses for Exchange Students: Engineering: Biomedical-Industrial-Mechanical - semester I b MSc Industrial Engineering and Management: Production Technology and Logistics  (Optional technical PTL modules) - semester I b Elective SSCA MSc Mechanical Engineering: Advanced Instrumentation  (Electives ) 1 semester I b keuze MSc Mechanical Engineering: Smart Factories  (Electives ) 1 semester I b keuze