Modellistica e Identificazione dei Processi Dinamici
Modeling and identification of dynamic processes Anna Maria Perdon
The student should have a good knowledge of the notions provided by basic courses in Mathematics and in Automatic Control. It 'also useful to know how to use programming tools
KNOWLEDGE AND UNDERSTANDING:
The aim of the course is to provide the student with basic theoretical knowledge and practical methods for developing mathematical models from experimental data, techniques for the identification and methodologies for the analysis of models properties. The student will acquire the basics notions on the construction of models for linear dynamic processes and on prediction methods for parameters estimation.CAPACITY TO APPLY KNOWLEDGE AND UNDERSTANDING:
Students will develop the skills needed to implement the methods and techniques learned in theory, they will learn to design a data collection experiment, to process them up to the identification of a suitable model and its validation. They will acquire the skills to work in the laboratory and to use dedicated software tools.TRANSVERSAL SKILLS:
Through guided exercises the students develop: the ability to learn by assessing the completeness and adequacy of their preparation; the independence of judgment in analysing the behaviour of economic and production processes and formulate and propose solutions to problems inherent in the representation and management of such models; the communication skills in formulating and properly describe the solutions to the problems under consideration.
Introduction and generalities about model construction and systems identification from experimental data. Models and parametric identification. Data collection and related problems. Best model and identification techniques (LS, ML, recursive methods). Model validation. Systems in state space form. Structural properties and relations between state space representations and external I/O representations. Realization of transfer function in state space form. Generalities on Neural Networks. An outline of identification by Neural Networks. Implementation of identification methods by Matlab System Identification Toolbox. Lab with NI myDAQ and NI myRIO, low-cost portable data acquisition (DAQ) devices that uses NI LabVIEW-based software instruments. Participation in laboratory activities is mandatory
Development of the examination
LEARNING EVALUATION METHODS
The learning evaluation will consist of a written test consisting consists of four questions of a theoretical nature, on the topics discussed in class and contained in the materials provided to the students. Each student must also complete a laboratory activity. report on the practical project on one of the topics discussed in class and present a report on this activity. In case of negative results of the written examination, the student can repeat it, during the same academic year.
LEARNING EVALUATION CRITERIA
Correctness, completeness and clarity in answering the questions in the theory test. Accuracy and completeness in solving the exercises. In the report on the laboratory activity the student must prove that he can apply the concepts learned in the course, to properly use the tools and appropriate technologies and to write a clear technical report.
LEARNING MEASUREMENT CRITERIA
The written test consists of 4 groups of questions on the various parts of the program, each group contains a question which is assigned a score between 0 and 10, and a question which is assigned a score between 0 and 6. The student must answer a question in each group, choosing two questions for 10 points and two for 6 points. The test is considered sufficient if the score is greater or equal to 15. The report on the laboratory activity is assigned a score from 0 to 30 and is sufficient only if the score is greater or equal to 18.
FINAL MARK ALLOCATION CRITERIA
The overall grade is given by the arithmetic mean, rounded up to the whole, of the sum of the scores obtained in the written test and in the report on the laboratory activity if both are sufficient. The overall grade required to pass the exam is 18 points. Otherwise the overall grade is Not sufficient . The student who in addition to getting a score greater than or equal to 30 has demonstrated complete mastery of the topics addressed, and clarity of exposition will have a 30 e lode.
Identificazione dei Modelli e Controllo Adattativi, S. Bittanti Pitagora Editrice Bologna
Slides and exercises can be found on the web site Esse3 Web
- Ingegneria Informatica e dell'Automazione (Corso di Laurea Triennale (DM 270/04))