A Hammerstein-Wiener Modeler already exists in the Systems Identification Toolbox in MATLAB. However, this has some decisive disadvantages, which ultimately led to the in-house development. The most significant point is the lack of support for MIMO models. The documentation of the MATLAB toolbox is mainly limited to the SISO case and tests with trivial examples suggest that the toolbox lacks proper support for MIMO models. For modelling RF devices, which in most cases have more than one port, this part of the System Identification Toolbox is only moderately suitable. What does work excellently, however, is the analysis of the linear dynamic part of the Hammerstein-Wiener model with MATLAB methods. In the SISO case, the ARX model used in our toolbox corresponds to a special IIR filter, the structure of which is given here.
The latter can be analyzed easily with the Digital Signal Processing Toolbox. For the MIMO case, however, the ARX model can also be converted into a state-space model. With the evaluation of its state transition matrix, the same is accomplished in the end. Therefore, no analysis methods are offered in the in-house Hammerstein-Wiener Modeler, as these are already sufficiently available. SISO and MIMO models are also fully supported there.