Data-driven control of biomedical systems

Standard approaches to control are based on first-principle models: they assume that some accurate first-principle model describing the dynamics of the plant to control is available. However, in many biomedical applications, deriving an accurate model is an extremely difficult task since the system dynamics is not well known and/or too complex. In this activity, a data-driven model predictive control (PMPC) approach for nonlinear systems is being developed, relying on the identification from data of polynomial prediction models. The main advantages of this approach over the standard methods are that it does not require a detailed knowledge of the plant to control and it is computationally efficient. Thanks to these features, the data-driven approach appears particularly suitable for the biomedical field, where deriving accurate physiological models can be extremely difficult. Applications include:

  • Regulation of blood glucose concentration in type 1 diabetic patients.
  • Control of cell populations finalized at tumor treatment.

ERC Sector:

  • PE7_1 Control engineering
  • PE7_3 Simulation engineering and modelling

Keywords:

  • Biomechatronics
  • Biomedical engineering
  • Data-driven control

Gruppi di ricerca

Persone di riferimento