Uncertainty Analysis

Compute parameter variability, plot confidence bounds

When you estimate the model parameters from data, you obtain their nominal values that are accurate within a confidence region. The size of this region is determined by the values of the parameter uncertainties computed during estimation. The magnitude of the uncertainties provide a measure of the reliability of the model. You can compute and visualize the effect of parameter uncertainties on the model response in time and frequency domains.

Functions

expand all

Simulation and Sampling

`simsd`	Simulate linear models with uncertainty using Monte Carlo method
`simsdOptions`	Option set for `simsd`
`freqresp`	Evaluate system response over a grid of frequencies
`rsample`	Random sampling of linear identified systems

Model Data

`present`	Display model information, including estimated uncertainty
`tfdata`	Access transfer function data
`zpkdata`	Access zero-pole-gain data

Covariance Data

`getcov`	Parameter covariance of identified model
`setcov`	Set parameter covariance data in identified model
`translatecov`	Translate parameter covariance across model transformation operations

Response Plots

`step`	Step response of dynamic system
`stepplot`	Plot step response of dynamic system
`impulse`	Impulse response plot of dynamic system; impulse response data
`bode`	Bode frequency response of dynamic system
`bodemag`	Magnitude-only Bode plot of frequency response
`nyquist`	Nyquist response of dynamic system
`nyquistplot`	Plot Nyquist response of dynamic system
`iopzmap`	Plot pole-zero map for input-output pairs of dynamic system using default options
`iopzplot`	Plot pole-zero map for input-output pairs of dynamic system
`showConfidence`	Display confidence regions on response plots for identified models

Topics

Compute Model Uncertainty
Compute model parameter uncertainty of linear models.
Plot Impulse and Step Response Using the System Identification App
To create a transient analysis plot in the System Identification app, select the Transient resp check box in the Model Views area.
Plot Bode Plots Using the System Identification App
To create a frequency-response plot for linear models in the System Identification app, select the Frequency resp check box in the Model Views area.
Plot the Noise Spectrum Using the System Identification App
To create a noise spectrum plot for parametric linear models in the app, select the Noise spectrum check box in the Model Views area.
Plot the Noise Spectrum at the Command Line
To plot the disturbance spectrum of an input-output model or the output spectrum of a time series model, use spectrum.
Model Poles and Zeros Using the System Identification App
To create a pole-zero plot for parametric linear models in the System Identification app, select the Zeros and poles check box in the Model Views area.