A major challenge associated with cardiac modelling is to use the large amount of data available on the system to circumvent the lack of an absolute ground truth for modelling, as each model should, in fact, becomes patient specific, with possibly nonstandard conditions associated with disease. Data assimilation, originally used in the environmental sciences, has now found its way into the life sciences and, in particular, cardiac modeling and aim at registering a mathematical model over available data. In this talk, we present a control-oriented data assimilation strategy called observer theory in which data-driven feedback expressions control the trajectory of the simulated system so that it pursues the real target trajectory and some parameters are jointly identified. The design of the feedback clearly depends on the type of model considered – mechanics, reaction diffusion, wave or transport systems – but is also closely related to the type of data considered. Unless we want to rely on approximate post-processing of the data, it is a major challenge to use the data as they are essentially supplied, which are mostly just images of the patterns or shapes of the underlying physical quantities. We will show examples of such observer design when we have access only to the front propagation of the cardiac action potential or map only the shape of the myocardial region that deforms during a heartbeat.
The seminar will be held in a hybrid format. We strongly encourage you to participate in person at MR 12 , Centre of Mathematical Sciences, CB3 0WA . Althernatively you can join using the following Zoom link:
- Speaker: Dr Philippe Moireau, Inria and Ecole Polytechnique
- Wednesday 25 January 2023, 13:00–14:00
- Venue: This seminar will be held in a hybrid format at MR 12, Centre of Mathematical Sciences, CB3 0WA. Alternatively you can also join with Zoom (see abstract for Zoom link)..
- Series: CMIH Hub seminar series; organiser: Yuan Huang.