Learn more about System Identification Toolbox

predict(idnlgrey) - Predict output k steps ahead for nonlinear ODE model

Syntax

YP = predict(NLSYS,DATA);
[YP,X0,XFINAL] = predict(NLSYS,DATA);
[YP,X0,XFINAL] = predict(NLSYS,DATA,K);
[YP,X0,XFINAL] = predict(NLSYS,DATA,K,X0INIT);

Description

YP = predict(NLSYS,DATA); predicts the k-step ahead output with an idnlgrey model.

[YP,X0,XFINAL] = predict(NLSYS,DATA); returns the initial states used in the prediction as well as the final states computed, in addition to the predicted output.

[YP,X0,XFINAL] = predict(NLSYS,DATA,K); specifies the prediction horizon to use during prediction.

[YP,X0,XFINAL] = predict(NLSYS,DATA,K,X0INIT); specifies the initialization for the k-step ahead prediction.

Input

• NLSYS: idnlgrey model for which output is to be predicted.

• DATA: Input-output data [Y U]. U is the input data that can be given either as an iddata object or as a matrix U = [U1 U2 ... Um], where the k:th column vector is input U_k. Similarly, Y is either an iddata object or a matrix of outputs (with as many columns as there are outputs). For time-continuous idnlgrey objects, DATA passed as a matrix will lead to that the data sample interval, Ts, is set to one.

• K: Prediction horizon. K and can be set to an integer between 1 and inf (pure simulation). As idnlgrey assumes an output error model structure, where prediction and simulation coincide, K has no meaning.

• X0INIT: Initial state to use. It can take the following values:

  • 'zero' : Zero initial state x(0) with all states fixed (nlsys.InitialStates.Fixed is thus ignored).

  • 'fixed': (or NLSYS.InitialState) Determines the values of the model initial states, but all states are fixed (NLSYS.InitialStates.Fixed is ignored).

  • 'estimate': NLSYS.InitialState determines the value of the initial states and all initial states are estimated (NLSYS.InitialStates.Fixed is ignored).

  • (Default) 'model': NLSYS.InitialState determines the value of the initial states, which initial states to estimate, and minimum and maximum state values.

  • vector/matrix: Column vector of initial states. For multiple-experiment DATA, x(0) can be a matrix where each column contains initial states for the corresponding experiment. All initial states are fixed (nlsys.InitialStates.Fixed is ignored).

  • struct array : An Nx-by-1 structure array with fields:

    • Name: Name of the state (a string).

    • Unit: Unit of the state (a string).

    • Value: Value of the states (a finite real 1-by-Ne vector, where Ne is the number of experiments.)

    • Minimum: Minimum values of the states (a real 1-by-Ne vector or a real scalar, in which case all initial states have the same minimum value).

    • Maximum: Maximum values of the states (a real 1-by-Ne vector or a real scalar, in which case all initial states have the same maximum value).

    • (Default) Fixed: Uses initial state values from NLSYS.InitialState.Value. Ignores NLSYS.InitialState.Fixed.

The initial state can also be specified using the property-value pair InitialState or X0, followed by the value X0INIT. For example: predict(NLSYS,DATA,K,'InitialState',X0INIT).

Output

• YP: Predicted output. If DATA is an iddata object, then YP will also be an iddata object. Otherwise, YP will be a matrix where the k:th output is found in the k:th column of YP. If DATA is a multiple experiment iddata object, then YP will be as well.

• X0: Initial states. In the single experiment case it is a column vector of length Nx. For multi-experiment data, X0 is an Nx-by-Ne matrix with the ith column specifying the initial state of experiment i.

• XFINAL: Final states computed. In the single experiment case it is a column vector of length Nx. For multi-experiment data, XFINAL is an Nx-by-Ne matrix with the ith column specifying the initial state of experiment i.

  Note   If predict is called without an output argument, MATLAB software displays the predicted output(s) in a plot window.

See Also

sim(idnlgrey)
pe
pem

predict(idnlarx)

predict(idnlhw)

© 1984-2009- The MathWorks, Inc.    -   Site Help   -   Patents   -   Trademarks   -   Privacy Policy   -   Preventing Piracy   -   RSS