Projects (Principal investigator)

Kernel-based regularization methods for system identification

(2015-2018, 3.6M SEK) funded by Swedish Research Council (VR)

Regularization is widely used in the area of statistics and machine learning for handling the bias-variance tradeoff and inducing sparse solutions. Recent results show that well designed and tuned regularization can outperform the standard ones for problems of model estimation and finding sparse solutions in system identification. However, such activities have not been particularly pronounced in the system identification community which has largely been sticking to the maximum likelihood (or related) framework. In this project, we intend to investigate how estimation of more sophisticated and practically useful models can be handled with regularization methods.

Projects (Participant)

LEARN (Limitations, Estimation, Adaptivity, Reinforcement, Networks in System Identification)

(2011-2015) an advanced grant under contract 267381 funded by European Research Council (ERC)

CADICS (Control, Autonomy, and Decision-making in Complex Systems)

(2008-2018) a Linneaus Research Center funded by Swedish Research Council (VR)

MOVIII (Modeling, Visualization and Information Integration)

(2006-2010) a Strategic Research Center funded by Swedish Foundation for Strategic Research (SSF)

Theory of input-to-state stability with restrictions for time-varying nonlinear systems and its applications

(No. 412006) funded by Hong Kong Research Grants Council (RGC)

Small gain theory based robust input-to-state stabilization of nonlinear systems with time-variant uncertainties

(No. 412305) funded by Hong Kong Research Grants Council (RGC)