Parameter Estimator

Estimate model parameters and initial states

Description

The Parameter Estimator app estimates parameters and initial states of a Simulink^® model using measured data. The app increases model accuracy so that the model reflects the measured hardware behavior. For example, you can automatically estimate electric motor resistance, inductance, and inertia from measured voltage and motor speed data.

Using this app, you can:

Import and preprocess measured data.
Find the most influential parameters to estimate (with the Sensitivity Analyzer).
Estimate model parameters and initial states, and monitor estimation progress.
Validate estimation results.

You can generate MATLAB^® code from the app, and accelerate parameter estimation using parallel computing and Simulink fast restart.

Open the Parameter Estimator App

Simulink Toolstrip: On the Apps tab, under Control Systems, click Parameter Estimator.
MATLAB command prompt: Enter spetool('modelname').

Examples

Programmatic Use

expand all

`spetool('modelname')`

spetool('modelname') opens the Parameter Estimator app and creates a new session. The model must be open or on the MATLAB path.

`spetool(spesession)`

spetool(spesession) opens a previously saved Parameter Estimator app session.

Version History

Introduced before R2006a

expand all

R2024b: Parameter estimation with Simscape

Starting in R2024b, you can directly select Simscape blocks and assign data in the Parameter Estimator app. Previously, you used a PS-Simulink Converter block to convert a physical signal into a Simulink output signal, selected the converted signal, and then assigned the data.

R2024a: Reference configuration set support

Starting in R2024a, Parameter Estimator supports models that use reference configuration sets.

R2023b: Simulink Online support

Starting in R2023b, Parameter Estimator is supported in Simulink Online™.

R2023a: Estimate discrete-valued parameters

Starting in R2023a, you can estimate discrete-valued parameters using Parameter Estimator. For more information, see Specify Parameters for Estimation.

R2022a: Surrogate optimization solver

Starting in R2022a, Parameter Estimator supports surrogate optimization to solve time-consuming optimization problems using potentially fewer objective function evaluations. For an example that uses surrogate optimization for response optimization, see Surrogate Optimization Using the Response Optimizer App.

R2020b: Name changed

The Parameter Estimation tool is now the Parameter Estimator app.

R2020b: Sequential quadratic programming algorithm improved

Because of improvements to the sequential quadratic programming algorithm, loading a session in Parameter Estimator can lead to different results if the session uses the Sequential Quadratic Programming option and was saved in a release from before R2016b.

For a session saved in a release older than R2016b, to force the app to use the improved algorithm, on the Parameter Estimation tab, click More Options and select Optimization Options. Then, in the Options dialog box, click OK.

R2019a: Export plots

Starting in R2019a, you can export plots from Parameter Estimation tool (Parameter Estimator) sessions. From a plot tab, use the Print menu to either print the plot or export the plot as a MATLAB figure.

R2018b: Support for model reference variables

Starting in R2018b, you can estimate parameters of Simulink model reference blocks using Parameter Estimation tool (Parameter Estimator).