Formal Verfication of Dynamical Systems using Integral Quadratic Constraints
Anders Rantzer
Abstract:
A grand challenge for verification of control systems could be stated as follows: Given requirement specifications, process model (including uncertainty description) and control software, prove that the control system will meet the specifications. Moreover, the verification should be scalable to problems of realistic size.
In this presentation, we explain how verification of dynamical systems can be done using Integral Quadratic Constraints. Such methods have a long history in control theory, but the issue of scalability has been a bottleneck until very recently. Here we explain the main ideas, including recent progress on scalability, and suggest challenges for further research.