checkPrediction

Compare outputs predicted by data-driven model to validation outputs

Since R2026a

collapse all in page

    Syntax

    checkPrediction(ddobj,Uval,Yval)

    Description

    checkPrediction(ddobj,Uval,Yval) supplies the input Uval to the prediction model of ddobj and compares the resulting output with the validation output data Yval.

    example

    Examples

    collapse all

    Open Live Script

    This example shows how to use the checkPrediction function to validate the prediction model of a DataDrivenMpc object.

    Create data-driven MPC controller

    Define the sample time and time vector.

    Ts = 0.1;
t = 0:Ts:10;

    Use a simple integrator as a plant to produce output data.

    plant = tf(1,[1 0]);

    Create model input and output data vector.

    dataU = (rand(length(t),1)-0.5)*3;
dataY = lsim(plant,dataU,t,0);

    Create a DataDrivenMPC object.

    ddobj = DataDrivenMPC(dataU,dataY,Ts);

    Define the main properties.

    ddobj.FutureSteps = 21; 
ddobj.PastSteps = 4;
ddobj.PlantOrder = 1;

    Check the controller prediction model

    Create the validation time vector. You must use a number of samples equal to the sum of the values of the FutureSteps and PastSteps properties.

    Nval = ddobj.FutureSteps + ddobj.PastSteps;
tval = Ts*(0:Nval-1)';

    Create the validation input and output data.

    dataUval = (rand(length(tval),1)-0.5)*3;
dataYval = lsim(plant,dataUval,tval,0);

    Use the checkPrediction function to validate the prediction model of ddobj.

    checkPrediction(ddobj,dataUval,dataYval)

    Figure contains an axes object. The axes object with title Prediction Validation, xlabel time, ylabel y(1) contains 2 objects of type line. One or more of the lines displays its values using only markers These objects represent measured output, predicted output.

    The plot shows how well the output of the predicted model (displayed with red circles) matches the validation output dataYval (displayed with a blue line).

    Input Arguments

    collapse all

    Data-driven model predictive controller, specified as a DataDrivenMPC object. To create a data-driven MPC controller, use DataDrivenMPC.

    Example: ddobj = DataDrivenMPC((0.5-rand(100,1))*2,lsim(tf(1,[1 1]),(0.5-rand(100,1))*2,0.1*(1:100)',0),0.1); creates the DataDrivenMPC object ddobj from the input signal (0.5-rand(100,1))*2, and the output signal obtained by supplying the input signal to the LTI plant tf(1,[1 1]), with a time vector 0.1*(1:100)', a sample time of 0.1 seconds, and an initial state of 0.

    Plant input validation data, specified as a matrix with as many columns as the number of plant inputs, and as many rows as the ddobj.Dimensions.HankelDepth property. The data must be finite and uniformly sampled with the sample time Ts, given as a third argument.

    Example: (0.5-rand(50,1))*2

    Plant output validation data, specified as a matrix with as many columns as the number of plant outputs, and as many rows as the ddobj.Dimensions.HankelDepth property. The data must be finite and uniformly sampled with the sample time Ts, given as a third argument.

    Example: lsim(ss(-1,1,1,0),(0.5-rand(50,1))*2,0.1*(1:50)',0)

    Tips

    If the quality of the prediction is not satisfactory, consider adjusting the PastSteps, FutureSteps, or PlantOrder property of ddobj or recreating ddobj using a different input-output trajectory.

    Algorithms

    This function divides the input and output validation data into past and future values (Up, Yp,Uf,Yf), and then uses the Hankel matrices in the ddobj object to estimate the future output Yefaccording to the method described in Data-Driven Open-Loop Output Prediction. The function then plots the estimated future output Yef against the validation future output Yf to show the quality of the prediction.

    Version History

    Introduced in R2026a

    See Also

    Functions

    Objects

    Blocks

    Topics