A bilinear extension of subspace identification for systems subject to white inputs

We generalize a class of existing linear subspace identification techniques to subspace identification algorithms for bilinear, discrete time, time invariant systems. A major assumption we make is that the input should be white. It is shown that in that case most of the properties of linear subspace identification theory can be extended to equivalent properties for bilinear systems. The practical interest of the method is illustrated by applying it to a model of a distillation column