Comparison of three methods for the identification of switched hybrid systems
This talk addresses the problem of parameter identification for switched ARX system. The identification of such systems typically results in non-convex optimization problems, where finding the globally optimal solution exhibits exponential computational complexity in the size of the input. The exponential complexity may however not be tractable even for middle size problems. Another approach involves heuristics in order to reduce the computational complexity, with the trade-off, that the estimates are only approximate solutions.
Three recently proposed open-source algorithms for switched ARX system identification are compared. We consider a modified k-means algorithm: k-LinReg, and two methods using sum-of-norms regularization: PWARX and son-em.
Statistical measures are introduced in order to quantitatively compare the performance of the different methods on a simulated one-dimensional example. The individual behavior of the methods on different generated systems is also analyzed.