A Robust Exact Differentiator Block for MATLAB/Simulink

We have implemented a differentiator block for Simulink. The implemented algorithm is based on the ideas of higher order sliding mode differentiators. They are known to exactly compute derivatives of signals and, in the case of noisy input signals appealing filtering options helping for generating estimates of time-derivatives are available.

Meanwhile the toolbox is available as version 1, version 2 and version 3. Latter is recommended to be used in simulations where the accuracy of the estimation is limited due to a large sampling step size. In version 3 an improved estimation accuracy may be the result for large simulation step sizes, filtering options are available and the source code is part of the package.

An observability normal form for linear, time-invariant and multiple-input multiple-output systems subject to disturbances

We have developed an algorithm for transforming linear, time-invariant and multiple-input multiple-output systems which are affected by external disturbances and are strong observable into new normal form. This normal form is especially helpful in the design of unknown input observers. The transformation algorithm is provided in the download section of this webpage.