Researchers: U. Jönsson, in cooperation with A. Megretski (M.I.T), A. Rantzer (Lund Institute of Technology, C. Kao (M.I.T), M. Rotea (Purdue University), and F. D'Amato (Purdue University).
Sponsor: The Swedish Research Council for Engineering Sciences (TFR).
We are involved in an effort to develop the framework of Integral Quadratic Constraints (IQCs). The notion integral quadratic constraint was introduced by Yakubovich in the 1960s and it has been shown to be a useful concept in systems analysis. Indeed, it can be used to unify much of the existing stability theory and it results in stability criteria that can be verified using convex optimization.
There has been a rapid development in this research area during the last few years. One important direction is the development efficient and user friendly software for IQC analysis. One of the first efforts was published in [A14]. This work has been continued in collaboration with with Megretskis group at LIDS-MIT, where the software package IQCbeta was developed [R12]. We have also worked on theoretical issues of IQC optimization. One contribution was to consider the quality of approximations to infinite dimensional IQC optimization using duality theory [A15]. Another was the development of a new algorithm for fast IQC optimization in [C28].
Another direction of research is the development of new IQC descriptions of system components and noise signals. We published two papers that discuss modeling of noise and disturbance signals using IQCs [C22], [C26]. We have also published several papers on analysis of uncertain systems with nonlinearities during this year. In [A16], we used an encapsulation technique to obtain general analysis results for a class of uncertain neutrally stable systems in feedback interconnection with slope restricted nonlinearities. Analysis of systems with multiple repeated slope restricted nonlinearities was discussed in [C25] using an IQC developed in [A9].
Research 1999/2000
Hybrid control of autonomous system
KTH optimization laboratory