Identification of LPV models for spatially varying interconnected systems

This paper presents a method for system identification of two dimensional parameter varying systems. The idea of linear parameter varying (LPV) identification for one dimensional causal systems in input-output form is extended to identify two dimensional (2-D) parameter varying systems in input-output form which may be causal, semi-causal or non-causal. The identification method is general in the sense that it can be used to identify systems which may be separable or non-separable. Furthermore, the algorithm explicitly takes boundary conditions into consideration. The effectiveness of the approach is demonstrated by applying it to identify a spatially varying model for a non-uniform beam.