Le jeudi 15 Septembre, nous avons accueilli Laurent Bako pour un séminaire d’axe:

Titre : Sparsity-inducing methods in system identification and state estimation

Résumé :
In this talk we will discuss a general sparsity-inducing optimization framework and its application potential for various engineering problems.
Indeed many control problems admit a mathematical formalization which naturally involves some structural sparsity constraints (e.g., network connectivity design/estimation, maximum hands-off control, sensor/actuator scheduling, control allocation, fault-tolerant control, switched systems identification, robust estimation under adversary attacks, etc). Sparsity-inducing optimization refers to a class of mathematical optimization problems in which the to-be-optimized objective function or the associated constraints involve an integer measure on the data such as for example the rank of a matrix or the cardinality of a set.
The talk will focus in particular on the application of the sparse optimization framework to system identification and state estimation.
We will discuss some striking properties of estimators designed from this framework: exact recovery capability, robustness and resilience.