Nonlinear State Monitoring With Karhunen-Loe`ve-Transform

Nonlinear dynamical systems often appear to have uncorrelated output. This characteristic leads to the idea of analyzing the dynamics of nonlinear systems with methods developed for random processes. The Karhunen-Loe`ve-Transform (KLT) was designed to detect coherent structures in random process data. It can also be applied for state monitoring of complex systems. This paper gives a short review on the mathematical concept of the KLT and discusses an approach to characterize the dynamics of nonlinear systems based on experimental data.