Learning of dynamical systems from a finite number of data points
Project Leader: Erik Weyer
Primary Contact: Erik Weyer (firstname.lastname@example.org)
Keywords: machine learning; mathematical modelling; signal processing; system identification
Disciplines: Electrical & Electronic Engineering
System identification deals with the problem of building mathematical models of dynamical systems from observed data. This is also known as data driven modelling or learning of dynamical systems. There will always be uncertainty associated with the obtained models, and in recent years methods have been developed which, instead of delivering a single model, deliver a set of model together with a ‘quality tag’ describing the probability that the true system belongs to the set. A particular feature of methods is that they do deliver guaranteed probabilities, which is a very attractive property.
There are several PhD projects available in the area of extending such methods to:
- Frequency domain modelling of dynamical systems
- Learning of dynamical networks
- Detecting changes or faults in dynamical systems
- Data driven control design.