Seminar Details
2013-11-28 (14.00) : Structure in Algorithmic Design for Identification and State Estimation
At Euler building
Organized by Mathematical Engineering
Speaker :
Michel Verhaegen (Delft Center for Systems and Control)
Abstract :
Mathematical models can be interpreted as structured relationship to represent (or approximate) physical dynamical phenomena. In this lecture we revisit how structure can be integrated in (1) subspace identification and (2) the solution of large scale receding horizon state reconstruction problems. In the first problem a solution is presented that integrates the separate algorithmic steps that so-far have characterized many subspace identification methods. It will be demonstrated that the integrative step enables to incorporate in an "elegant" manner structural information about the underlying system to be identified. A convex relaxation of the integrative solution is given. For the second problem the receding horizon estimation problem is used as illustration of a framework to design distributed algorithmic solutions for large scale systems and control problems. The lecture is concluded with a brief summary of the ERC project iCON.