005. Parameter-varying models

In data-driven modelling, one can estimate linear time-varying models from measurements [1]. In many situations, the time variation is linked to some observable signals such as the temperature, the varying set point, the State-of-Charge (SoC) of a battery or the trajectory of a robot-arm. These signals are called scheduling parameters. In a first step a time-varying model is constructed with existing techniques (see figure). The challenge is to construct a parameter-varying model from this initial result. Applications of parameter-varying models can, for example, be found in the control of nonlinear systems.

Linear time

[1] Hallemans, N., Pintelon, R., Zhu, X., Collet, T., Claessens, R., Wouters, B., Hubin, A. and Lataire, J., 2020. Detection, Classification, and Quantification of Nonlinear Distortions in Time-Varying Frequency Response Function Measurements. IEEE Transactions on Instrumentation and Measurement70, pp.1-14.


Prerequisites: good knowledge in system theory, signal processing, measurement, data-based modelling and Matlab.



Back to top