Identification of LPV state-space models using ℋ 2- minimisation

Abstract

Advanced robustness analysis methods employed in flight control law clearance such as IQC-analysis and μ-analysis, rely on linear fractional representations (LFRs). These models are usually obtained from linear parameter varying (LPV)-models which approximate the behaviour of the underlying parameter uncertain nonlinear aircraft model. The generation of LPV-models is usually done starting from a collection of linearised state-space models describing the local behaviour of the nonlinear aircraft for a representative set of parameter values and flight conditions. In this chapter we propose an optimisation-based generation method to convert these linear models into an LPV-model, by minimising a suitable ℋ 2 error norm. Although computationally more demanding than the alternative element-wise approximation approach, the new method often produces LPV-models with lower complexity. © 2012 Springer-Verlag Berlin Heidelberg.