Recursive Polynomial Model Estimator

Estimate input-output and time-series polynomial model coefficients

expand all in page
  • Library:

  • System Identification Toolbox / Estimators

  • Recursive Polynomial Model Estimator block

Description

Model Structures

Use the Recursive Polynomial Model Estimator block to estimate discrete-time input-output polynomial and time-series models.

These model structures are:

  • AR — A(q)y(t) = e(t)

  • ARMA — A(q)y(t) = C(q)e(t)

  • ARX — A(q)y(t) = B(q)u(t – nk) + e(t)

  • ARMAX — A(q)y(t) = B(q)u(t – nk) + C(q)e(t)

  • OE — y(t)=B(q)F(q)u(t−nk)+e(t)

  • BJ — y(t)=B(q)F(q)u(t−nk)+​C(q)D(q)e(t)

q is the time-shift operator and nk is the input delay. u(t) is the input, y(t) is the output, and e(t) is the error. For MISO models, there are as many B(q) polynomials as the number of inputs.

The orders of these models correspond to the maximum number of time shifts, as represented by the exponent of q. For instance, the order na is represented in the A(q) polynomial by:

1 + a1q-1 + a2q-2 + … + anaq-na.

An equivalent representation applies to the C(q), D(q), and F(q) polynomials and their corresponding orders nc, nd, and nf.

The B(q) polynomial is unique with respect to the others, because this polynomial operates on the input and contains the system zeros. For B(q), the order nb is the order of the polynomial B(q) + 1:

b1 + b2q-1 + b3q-2 + … + bnbq-(nb-1).

The orders na, nb, nc, nd, nf, and input delay nk are known ahead of time. Specify these values as block parameters. Provide u(t) and y(t) through the Inputs and Outputs inports, respectively. The block estimates the set of A(q), B(q), C(q), D(q), and F(q) coefficients that the model structure uses and outputs them in the Parameters outport. During the estimation, the block constrains the estimated C, D, and F polynomials to a stable region with roots in the unit disk, while allowing the estimated A and B polynomials to be unstable. The Parameters outport provides a bus signal with the following elements:

  • A — Vector containing [1 a1(t) ... ana(t)].

  • B — Vector containing [01 … 0nk, b1(t) ... bnb(t)]. For MISO data, B is a matrix where the i-th row parameters correspond to the i-th input.

  • C — Vector containing [1 c1(t) ... cnc(t)].

  • D — Vector containing [1 d1(t) ... dnd(t)].

  • F — Vector containing [1 f1(t) ... fnf(t)].

For example, suppose that you want to estimate the coefficients for the following SISO ARMAX model:

y(t) + a1y(t – 1) +...+anay(t – na) = b1u(t – nk) + … + bnbu(t – nb – nk + 1) + e(t) + c1e(t – 1) + … + cnce(t – nc)

y, u, na, nb, nc, and nk are known quantities that you provide to the block. For each time step, t, the block estimates the A, B, and C parameter values, constraining only the C polynomial to a stable region. The block then outputs these estimated values using the Parameters outport.

Block Capabilities

The block supports several estimation methods and data input formats. The block can provide both infinite-history [1] and finite-history [2] (also known as sliding-window) estimates for θ. Configurable options in the block include:

  • Multiple inputs (ARX model structure only) — See the Inputs port.

  • Sample-based or frame-based data format — See the Input Processing parameter.

  • Multiple infinite-history estimation methods [1] — See the Estimation Method parameter.

  • Infinite-history (all model structures) or finite-history (AR, ARX, or OE model structures only) — See the History parameter.

  • Initial conditions, enable flag, and reset trigger — See the Initial Estimate, Add enable port, and External Reset parameters.

For more information on the estimation methods, see Recursive Algorithms for Online Parameter Estimation.

Ports

Input

expand all

Input signal u(t). The Input Processing parameter and the number of inputs nu define the dimensions of the signal. Only the ARX model structure can have multiple inputs, with nu greater than 1.

  • Sample-based input processing and nu inputs — nu-by-1 vector

  • Frame-based input processing with M samples per frame and nu inputs — M-by-nu matrix

Dependencies

To enable this port, set the Model Structure parameter to ARX, ARMAX, BJ, or OE.

Data Types: single | double

Measured output signal y(t). The Input Processing parameter defines the dimensions of the signal:

  • Sample-based input processing — Scalar

  • Frame-based input processing with M samples per frame — M-by-1 vector

Data Types: single | double

External signal that allows you to enable and disable estimation updates. If the signal value is:

  • true — Estimate and output the parameter values for the time step.

  • false — Do not estimate the parameter values, and output the most recent previously estimated value.

Dependencies

To enable this port, select the Add enable port parameter.

Data Types: single | double | Boolean | int8 | int16 | int32 | int64 | uint8 | uint16 | uint32

Reset parameter estimation to its initial conditions. The value of the External reset parameter determines the trigger type. The trigger type dictates whether the reset occurs on a signal that is rising, falling, either rising or falling, level, or on level hold.

Dependencies

To enable this port, select any option other than None in the External reset dropdown.

Data Types: single | double | Boolean | int8 | int16 | int32 | uint8 | uint16 | uint32

Initial parameter estimates, supplied from a source external to the block. The block uses this inport at the beginning of the simulation or when you trigger an algorithm reset using the Reset signal.

For information on the contents of the InitialParameters bus object, see the Parameters port description.

Dependencies

To enable this port, set History to Infinite and Initial Estimate to External.

Data Types: single | double

Initial parameter covariances, supplied from a source external to the block. For details, see the Parameter Covariance Matrix parameter. The block uses this inport at the beginning of the simulation or when you trigger an algorithm reset using the Reset signal.

Dependencies

To enable this port, set the following parameters:

  • History to Infinite

  • Estimation Method to Forgetting Factor or Kalman Filter

  • Initial Estimate to External

Data Types: single | double

Initial set of inputs when using finite-history (sliding-window) estimation, supplied from an external source.

  • If Model Structure is ARX, then the signal to this port must be a (W–1+max(nb)+max(nk))-by-nu matrix, where W is the window length and nu is the number of inputs. nb is the vector of B(q) polynomial orders and nk is the vector of input delays.

  • If Model Structure is OE, then the signal to this port must be a (W–1+nb+nk)-by-1 vector, where W is the window length. nb is the vector of B(q) polynomial orders and nk is the vector of input delays.

The block uses this inport at the beginning of the simulation or whenever the Reset signal triggers.

If the initial buffer is set to 0 or does not contain enough information, you see a warning message during the initial phase of your estimation. The warning should clear after a few cycles. The number of cycles it takes for sufficient information to be buffered depends upon the order of your polynomials and your input delays. If the warning persists, you should evaluate the content of your signals.

Dependencies

To enable this port, set:

  • History to Finite

  • Model Structure to ARX or OE

  • Initial Estimate to External

Data Types: single | double

Initial set of output measurements when using finite-history (sliding-window) estimation, supplied from an external source.

  • If Model Structure is AR or ARX, then the signal to this port must be a (W+na)-by-1 vector, where W is the window length and na is the polynomial order of A(q).

  • If Model Structure is OE, then the signal to this port must be a (W+nf)-by-1 vector, where W is the window length and nf is the polynomial order of F(q).

The block uses this inport at the beginning of the simulation or whenever the Reset signal triggers.

If the initial buffer is set to 0 or does not contain enough information, you see a warning message during the initial phase of your estimation. The warning should clear after a few cycles. The number of cycles it takes for sufficient information to be buffered depends upon the order of your polynomials and your input delays. If the warning persists, you should evaluate the content of your signals.

Dependencies

To enable this port, set:

  • History to Finite

  • Model Structure to AR, ARX, or OE

  • Initial Estimate to External

Data Types: single | double

Output

expand all

Estimated polynomial coefficients, returned as a bus. The bus contains an element for each of the A, B, C, D, and F polynomials that correspond to the structure that you specify in Model Structure (see Model Structures ).

Each bus element is a vector signal containing the associated polynomial coefficients. For example, the A element contains [1 a1(t) ... ana(t)].

Estimated C, D, and F values are constrained to be stable discrete-time polynomials. That is, these polynomials all have roots within the unit circle. Estimated A and B polynomials are allowed to be unstable.

Data Types: single | double

Estimation error, returned as:

  • Scalar — Sample-based input processing

  • M-by-1 vector — Frame-based input processing with M samples per frame

Dependencies

To enable this port, select the Output estimation error parameter.

Data Types: single | double

Parameter estimation error covariance P, returned as an N-by-N matrix, where N is the number of parameters. For details, see the Output Parameter Covariance Matrix parameter.

Dependencies

To enable this port:

  • Select the Output parameter covariance matrix parameter.

  • If History is Infinite, set Estimation Method to Forgetting Factor or Kalman Filter.

Data Types: single | double

Parameters

expand all

Model Structure

Estimated model structure, specified as one of the following:

  • ARX — SISO or MISO ARX model

  • ARMAX — SISO ARMAX model

  • OE — SISO OE model

  • BJ — SISO BJ model

  • AR — Time-series AR model

  • ARMA — Time-series ARMA model

Model Parameters

Specify how to provide initial parameter estimates to the block:

  • None — Do not specify initial estimates.

    The block uses 0 as the initial parameter estimate.

    Specify the parameters that the block enables based on your choice of model structure and estimation method.

    • Specify the set of Number of Parameters () parameters that the block enables based on your Model Structure. For instance, if your setting for Model Structure is AR, specify the Number of Parameters in A(q) (na) parameter.

    • Specify the Input Delay (nk) parameter that the block enables when your model structure uses a B(q) element.

    • Specify the Parameter Covariance Matrix if Estimation Method is Forgetting Factor or Kalman Filter.

  • Internal — Specify initial parameter estimates internally to the block.

    • Specify the initial parameter values Initial () parameters that the block enables based on your Model Structure and History. For instance, if your setting for Model Structure is AR and History is Infinite, specify the Initial A(q) parameter.

    • Specify the Input Delay (nk) parameter that the block enables when your model structure uses a B(q) element.

    • Specify the Parameter Covariance Matrix parameter if Estimation Method is Forgetting Factor or Kalman Filter.

    • Specify the Initial Inputs parameter (ARX and OE only) and the Initial Outputs parameter (ARX, AR, and OE) if History is Finite.

  • External — Specify initial parameter estimates as an input signal to the block.

    Specify the Number of Parameters () parameters that the block enables based on your Model Structure. Your setting for Model Structure and for the History parameter determines which signals to connect to the relevant ports:

    • If History is Infinite — InitialParameters and InitialCovariance

    • If History is Finite — InitialOutputs for the AR, ARX, and OE model structures, and InitialInputs for the ARX and OE model structures

Programmatic Use

Block Parameter: InitialEstimateSource
Type: character vector, string
Values: 'None', 'Internal', 'External'
Default: 'None'

Specify the number of estimated parameters na in the A(q) polynomial.

Dependencies

To enable this parameter, either:

  • Set History to Infinite, Model Structure to AR, ARX, ARMA, or ARMAX, and Initial Estimate to None or External.

  • Set History to Finite, Model Structure to AR or ARX, and Initial Estimate to None or External.

Programmatic Use

Block Parameter: A0
Type: non-negative integer
Default: 1

Specify the number of estimated parameters nb in the B(q) polynomial.

For MISO systems using an ARX model structure, specify nb as a vector with as many elements as there are inputs. Each element of this vector represents the order of the B(q) polynomial associated with the corresponding input. For example, suppose that you have a two-input MISO system whose B(q) elements are: [B1B2]=[0b1100b21b22]. The zero at the beginning of each polynomial represents a single input delay for each input (see the Initial B(q) parameter description). The trailing zero in B1 is for equalizing the length of the polynomials and has no impact on estimation. nb for each polynomial is equal to the number of estimated parameters following the initial zero, or 1 for input 1 and 2 for input 2. Specify Number of Parameters in B(q) (nb) as [1 2], and Input Delay (nk) as [1 1].

Dependencies

To enable this parameter, either:

  • Set History to Infinite, Model Structure to ARX, ARMAX, BJ, or OE, and Initial Estimate to None or External.

  • Set History to Finite with Model Structure of ARX or OE and Initial Estimate to None or External.

Programmatic Use

Block Parameter: B0
Type: positive integer
Default: 1

Specify the number of estimated parameters nc in the C(q) polynomial.

Dependencies

To enable this parameter, set History to Infinite, Model Structure to ARMA, ARMAX, or BJ, and Initial Estimate to None or External

Programmatic Use

Block Parameter: C0
Type: positive integer
Default: 1

Specify the number of estimated parameters nd in the D(q) polynomial.

Dependencies

To enable this parameter, set History to Infinite, Model Structure to BJ, and Initial Estimate to None or External.

Programmatic Use

Block Parameter: D0
Type: positive integer
Default: 1

Specify the number of estimated parameters nf in the F(q) polynomial.

Dependencies

To enable this parameter, set Initial Estimate to None or External and either:

  • History to Infinite, Model Structure to OE or BJ, and Initial Estimate to None or External

  • History to Finite, Model Structure to OE, and Initial Estimate to None or External

.

Programmatic Use

Block Parameter: F0
Type: positive integer
Default: 1

Specify the input delay as an integer representing the number of time steps that occur before the input affects the output. This delay is also called the dead time in the system. The block encodes the input delay as fixed leading zeros of the B(q) polynomial. For more information, see the B(q) parameter description.

For MISO systems with ARX model structure, specify nk as a vector with elements specifying the delay for each input. This vector is of length nu, where nu is the number of inputs.

For example, suppose that you have a two-input MISO system whose B(q) elements are: [B1B2]=[0b11b1200b21]. The zeros at the beginning of the polynomials represent a single-sample delay for the first input, and a two-sample delay for the second input (see the Initial B(q) parameter description). nb for each polynomial is equal to the number of estimated parameters following the initial zeros, or 2 for input 1 and 1 for input 2. Specify Input Delay (nk) as [1 2], and Number of Parameters in B(q) (nb) as [2 1].

Dependencies

To enable this parameter, either:

  • Set History to Infinite, Model Structure to ARX, ARMAX, OE, or BJ, and Initial Estimate to None or External.

  • Set History to Finite, Model Structure to ARX or OE, and Initial Estimate to None or External.

Programmatic Use

Block Parameter: nk
Type: non-negative integer vector
Default: 1

  • Real positive scalar, α — Covariance matrix is an N-by-N diagonal matrix, with α as the diagonal elements.

  • Vector of real positive scalars, [α(a),α(b),α(c), α(d), α(f)] — Covariance matrix is an N-by-N diagonal matrix, with [α(a),α(b),α(c),α(d), α(f)] as the diagonal elements. α(a) is a vector of the covariance for each coefficient of the A polynomial. Similarly, α(b), α(c), α(d) and α(f) are vectors containing the covariance of the coefficients of the B, C, D and F polynomials, respectively.

  • N-by-N symmetric positive-definite matrix.

    N can be one of the following:

    • AR — N = na

    • ARX — N​=na+∑i=1Nunbi

    • ARMA — N = na + nc

    • ARMAX — N = na + nb + nc

    • OE — N = nb + nf

    • BJ — N = nb + nc + nd + nf

Dependencies

To enable this parameter, set

  • History to Infinite

  • Initial Estimate to None or Internal

  • Estimation Method to Forgetting Factor or Kalman Filter

The block uses this parameter at the beginning of the simulation or whenever the Reset signal triggers.

Programmatic Use

Block Parameter: P0
Type: scalar, vector, or matrix
Default: 1e4

Specify the initial estimate of the A(q) polynomial coefficients as a row vector of length na+1.

The leading coefficient of A must be 1.

Dependencies

To enable this parameter, set:

  • Model Structure to AR, ARX, ARMA, or ARMAX

  • Initial Estimate to Internal

Programmatic Use

Block Parameter: A0
Type: real vector
Default: [1 eps]

Specify the initial estimate of the B(q) polynomial coefficients as a row vector of length nb+nk. For multiple-input models, specify a matrix where the ith row corresponds to the ith input.

The block counts the leading zeros in B(q) and interprets them as input delay nk. Those zeros are fixed throughout the estimation. nb is the number of elements after the first nonzero element in B(q). The block estimates the value of these nb elements.

For example:

  • [0 eps] corresponds to nk=1 and nb=1.

  • [0 0 eps] corresponds to nk=2 and nb=1.

  • [0 0 eps 0 eps] corresponds to nk=2 and nb=3.

The default value is [0 eps].

Dependencies

To enable this parameter, set:

  • Model Structure to ARX, ARMAX, OE, or BJ

  • Initial Estimate to Internal

Programmatic Use

Block Parameter: B0
Type: real vector or matrix
Default: [0 eps]

Specify the initial estimate of the C(q) polynomial coefficients as a row vector of length nc+1.

The leading coefficient of C(q) must be 1.

The coefficients must define a stable discrete-time polynomial, that is, have all polynomial roots within the unit circle.

Dependencies

To enable this parameter, set:

  • History to Infinite

  • Model Structure to ARMA, ARMAX, or BJ

  • Initial Estimate to Internal

Programmatic Use

Block Parameter: C0
Type: real vector
Default: [1 eps]

Specify the initial estimate of the D(q) polynomial coefficients as a row vector of length nd+1.

The leading coefficient of D(q) must be 1.

The coefficients must define a stable discrete-time polynomial, that is, have all polynomial roots within the unit circle.

Dependencies

To enable this parameter, set:

  • History to Infinite

  • Model Structure to BJ

  • Initial Estimate to Internal

Programmatic Use

Block Parameter: D0
Type: real vector
Default: [1 eps]

Specify the initial estimate of the F(q) polynomial coefficients as a row vector of length nf+1.

The leading coefficient of F(q) must be 1.

The coefficients must define a stable discrete-time polynomial, that is, have all polynomial roots within the unit circle.

Dependencies

To enable this parameter, set:

  • Model Structure to OE or to BJ

  • Initial Estimate to Internal

Programmatic Use

Block Parameter: F0
Type: real vector
Default: [1 eps]

Specify initial values of the inputs buffer when using finite-history (sliding window) estimation. The buffer dimensions accommodate the specified window length, the regressors associated with polynomials within that window, the input delays, and the number of inputs. These elements drive a matrix size of:

  • ARX model structure — (W–1+max(nb)+max(nk))-by-nu

  • OE model structure — (W–1+nb+nk)-by-1

where W is the window length and nu is the number of inputs. nb is the vector of B(q) polynomial orders and nk is the vector of input delays.

When the initial value is set to 0, the block populates the buffer with zeros.

If the initial buffer is set to 0 or does not contain enough information, you see a warning message during the initial phase of your estimation. The warning should clear after a few cycles. The number of cycles it takes for sufficient information to be buffered depends upon the order of your polynomials and your input delays. If the warning persists, you should evaluate the content of your signals.

The block uses this parameter at the beginning of the simulation or whenever the Reset signal triggers.

Dependencies

To enable this parameter, set

  • History to Finite

  • Model Structure to ARX or OE

  • Initial Estimate to Internal

.

Programmatic Use

Block Parameter: InitialInputs
Type: real matrix
Default: 0

Specify initial values of the measured outputs buffer when using finite-history (sliding-window) estimation. The buffer dimensions accommodate the specified window length and the regressors associated with the polynomials within that window.

  • AR or ARX model structure — (W+na)-by-1 vector, where W is the window length and na is the polynomial order of A(q).

  • OE model structure — (W+nf)-by-1 vector, where W is the window length and nf is the polynomial order of F(q).

When the initial value is set to 0, the block populates the buffer with zeros.

If the initial buffer is set to 0 or does not contain enough information, you see a warning message during the initial phase of your estimation. The warning should clear after a few cycles. The number of cycles it takes for sufficient information to be buffered depends upon the order of your polynomials and your input delays. If the warning persists, you should evaluate the content of your signals.

The block uses this parameter at the beginning of the simulation or whenever the Reset signal triggers.

Dependencies

To enable this parameter, set:

  • History to Finite

  • Model Structure to AR, ARX, or OE

  • Initial Estimate to Internal

Programmatic Use

Block Parameter: InitialOutputs
Type: real vector
Default: 0

Input Processing and Sample Time

  • Sample-based processing operates on signals streamed one sample at a time.

  • Frame-based processing operates on signals containing samples from multiple time steps. Many machine sensor interfaces package multiple samples and transmit these samples together in frames. Frame-based processing allows you to input this data directly without having to first unpack it.

Specifying frame-based data adds an extra dimension of M to some of your data inports and outports, where M is the number of time steps in a frame. These ports are:

  • Inputs

  • Output

  • Error

For more information, see the port descriptions in Ports.

Programmatic Use

Block Parameter: InputProcessing
Type: character vector, string
Values: 'Sample-based', 'Frame-based'
Default: 'Sample-based'

Specify the data sample time, whether by individual samples for sample-based processing (ts), or by frames for frame-based processing (tf = Mts), where M is the frame length. When you set Sample Time to its default value of –1, the block inherits its ts or tf based on the signal.

Specify Sample Time as a positive scalar to override the inheritance.

Programmatic Use

Block Parameter: Ts
Type: real scalar
Values: –1, positive scalar
Default: –1

Algorithm and Block Options

Algorithm Options

The History parameter determines what type of recursive algorithm you use:

  • Infinite — Algorithms in this category aim to produce parameter estimates that explain all data since the start of the simulation. These algorithms retain the history in a data summary. The block maintains this summary within a fixed amount of memory that does not grow over time.

    The block provides multiple algorithms of the Infinite type. Selecting this option enables the Estimation Method parameter with which you specify the algorithm.

  • Finite — Algorithms in this category aim to produce parameter estimates that explain only a finite number of past data samples. The block uses all of the data within a finite window, and discards data once that data is no longer within the window bounds. This method is also called sliding-window estimation.

    The block provides one algorithm of the Finite type. You can use this option only with the AR, ARX, and OE model structures.

    Selecting this option enables the Window Length parameter.

For more information on recursive estimation methods, see Recursive Algorithms for Online Parameter Estimation

Programmatic Use

Block Parameter: History
Type: character vector, string
Values: 'Infinite', 'Finite'
Default: 'Infinite'

The Window Length parameter determines the number of time samples to use for the finite-history (sliding-window) estimation method. Choose a window size that balances estimation performance with computational and memory burden. Sizing factors include the number and time variance of the parameters in your model. Always specify Window Length in samples, even if you are using frame-based input processing.

Window Length must be greater than or equal to the number of estimated parameters.

Suitable window length is independent of whether you are using sample-based or frame-based input processing. However, when using frame-based processing, Window Length must be greater than or equal to the number of samples (time steps) contained in the frame.

Dependencies

To enable this parameter, set History to Finite.

Programmatic Use

Block Parameter: WindowLength
Type: positive integer
Default: 200

Specify the estimation algorithm when performing infinite-history estimation. When you select any of these methods, the block enables additional related parameters.

Forgetting factor and Kalman filter algorithms are more computationally intensive than gradient and normalized gradient methods. However, these more intensive methods have better convergence properties than the gradient methods. For more information about these algorithms, see Recursive Algorithms for Online Parameter Estimation.

Programmatic Use

Block Parameter: EstimationMethod
Type: character vector, string
Values: 'Forgetting Factor','Kalman Filter','Normalized Gradient','Gradient'
Default: 'Forgetting Factor'

The forgetting factor λ specifies if and how much old data is discounted in the estimation. Suppose that the system remains approximately constant over T0 samples. You can choose λ such that:

T0=11−λ

  • Setting λ = 1 corresponds to “no forgetting” and estimating constant coefficients.

  • Setting λ < 1 implies that past measurements are less significant for parameter estimation and can be “forgotten.” Set λ < 1 to estimate time-varying coefficients.

Typical choices of λ are in the [0.98 0.995] range.

Dependencies

To enable this parameter, set History to Infinite and Estimation Method to Forgetting Factor.

Programmatic Use

Block Parameter: AdaptationParameter
Type: scalar
Values: (0 1] range
Default: 1

Process Noise Covariance prescribes the elements and structure of the noise covariance matrix for the Kalman filter estimation. Using N as the number of parameters to estimate, specify the Process Noise Covariance as one of the following:

  • Real nonnegative scalar, α — Covariance matrix is an N-by-N diagonal matrix, with α as the diagonal elements.

  • Vector of real nonnegative scalars, [α1,...,αN] — Covariance matrix is an N-by-N diagonal matrix, with [α1,...,αN] as the diagonal elements.

  • N-by-N symmetric positive semidefinite matrix.

The Kalman filter algorithm treats the parameters as states of a dynamic system and estimates these parameters using a Kalman filter. Process Noise Covariance is the covariance of the process noise acting on these parameters. Zero values in the noise covariance matrix correspond to constant coefficients, or parameters. Values larger than 0 correspond to time-varying parameters. Use large values for rapidly changing parameters. However, expect the larger values to result in noisier parameter estimates. The default value is 1.

Dependencies

To enable this parameter, set History to Infinite and Estimation Method to Kalman Filter.

Programmatic Use

Block Parameter: AdaptationParameter
Type: scalar, vector, matrix
Default: 1

The adaptation gain γ scales the influence of new measurement data on the estimation results for the gradient and normalized gradient methods. When your measurements are trustworthy, or in other words have a high signal-to-noise ratio, specify a larger value for γ. However, setting γ too high can cause the parameter estimates to diverge. This divergence is possible even if the measurements are noise free.

When Estimation Method is NormalizedGradient, Adaptation Gain should be less than 2. With either gradient method, if errors are growing in time (in other words, estimation is diverging), or parameter estimates are jumping around frequently, consider reducing Adaptation Gain.

Dependencies

To enable this parameter, set History to Infinite and Estimation Method to Normalized Gradient or to Gradient.

Programmatic Use

Block Parameter: AdaptationParameter
Type: scalar
Default: 1

The normalized gradient algorithm scales the adaptation gain at each step by the square of the two-norm of the gradient vector. If the gradient is close to zero, the near-zero denominator can cause jumps in the estimated parameters. Normalization Bias is the term introduced to the denominator to prevent these jumps. Increase Normalization Bias if you observe jumps in estimated parameters.

Dependencies

To enable this parameter, set History to Infinite and Estimation Method to Normalized Gradient.

Programmatic Use

Block Parameter: NormalizationBias
Type: scalar
Default: eps
Block Options

Use the Error outport signal to validate the estimation. For a given time step t, the estimation error e(t) is calculated as:

e(t)=y(t)−yest(t),

where y(t) is the measured output that you provide, and yest(t) is the estimated output using the regressors H(t) and parameter estimates θ(t-1).

Programmatic Use

Block Parameter: OutputError
Type: character vector, string
Values: 'off','on',
Default: 'off'

Use the Covariance outport signal to examine parameter estimation uncertainty. The software computes parameter covariance P assuming that the residuals, e(t), are white noise, and the variance of these residuals is 1.

The interpretation of P depends on the estimation approach you specify in History and Estimation Method as follows:

  • If History is Infinite, then your Estimation Method selection results in:

    • Forgetting Factor — (R2/2)P is approximately equal to the covariance matrix of the estimated parameters, where R2 is the true variance of the residuals. The block returns these residuals through the Error port.

    • Kalman Filter — R2P is the covariance matrix of the estimated parameters, and R1 /R2 is the covariance matrix of the parameter changes. Here, R1 is the covariance matrix that you specify in Parameter Covariance Matrix.

    • Normalized Gradient or Gradient — Covariance P is not available.

  • If History is Finite (sliding-window estimation) — R2P is the covariance of the estimated parameters. The sliding-window algorithm does not use this covariance in the parameter-estimation process. However, the algorithm does compute the covariance for output so that you can use it for statistical evaluation.

Programmatic Use

Block Parameter: OutputP
Type: character vector, string
Values: 'off','on'
Default: 'off'

Use the Enable signal to provide a control signal that enables or disables parameter estimation. The block estimates the parameter values for each time step that parameter estimation is enabled. If you disable parameter estimation at a given step, t, then the software does not update the parameters for that time step. Instead, the block output contains the last estimated parameter values.

You can use this option, for example, when or if:

  • Your regressors or output signal become too noisy, or do not contain information at some time steps

  • Your system enters a mode where the parameter values do not change in time

Programmatic Use

Block Parameter: AddEnablePort
Type: character vector, string
Values: 'off','on'
Default: 'off'

Set the External reset parameter to both add a Reset inport and specify the inport signal condition that triggers a reset of algorithm states to their specified initial values. Reset the estimation, for example, if parameter covariance is becoming too large because of lack of either sufficient excitation or information in the measured signals. The External reset parameter determines the timing for the reset.

Suppose that you reset the block at a time step, t. If the block is enabled at t, the software uses the initial parameter values specified in Initial Estimate to estimate the parameter values. In other words, at t, the block performs a parameter update using the initial estimate and the current values of the inports.

If the block is disabled at t and you reset the block, the block output contains the values specified in Initial Estimate.

Specify this option as one of the following:

  • None — Algorithm states and estimated parameters are not reset.

  • Rising — Trigger reset when the control signal rises from a negative or zero value to a positive value. If the initial value is negative, rising to zero triggers reset.

  • Falling — Trigger reset when the control signal falls from a positive or a zero value to a negative value. If the initial value is positive, falling to zero triggers reset.

  • Either — Trigger reset when the control signal is either rising or falling.

  • Level — Trigger reset in either of these cases:

    • Control signal is nonzero at the current time step.

    • Control signal changes from nonzero at the previous time step to zero at the current time step.

  • Level hold — Trigger reset when the control signal is nonzero at the current time step.

When you choose any option other than None, the software adds a Reset inport to the block. You provide the reset control input signal to this inport.

Programmatic Use

Block Parameter: ExternalReset
Type: character vector, string
Values: 'None','Rising','Falling', 'Either', 'Level', 'Level hold'
Default: 'None'

Model Examples

Online Recursive Least Squares Estimation

Online Recursive Least Squares Estimation

Implement an online recursive least squares estimator. You estimate a nonlinear model of an internal combustion engine and use recursive least squares to detect changes in engine inertia.

Estimate Parameters of System Using Simulink Recursive Estimator Block

Estimate Parameters of System Using Simulink Recursive Estimator Block

Use a model containing Simulink recursive estimator to accept input and output signals, construct a regressor signal, and estimate system parameters.

Use Frame-Based Data for Recursive Estimation in Simulink

Use Frame-Based Data for Recursive Estimation in Simulink

Use frame-based signals in a Simulink recursive estimation model.

References

[1] Ljung, L. System Identification: Theory for the User. Upper Saddle River, NJ: Prentice-Hall PTR, 1999, pp. 363–369.

[2] Zhang, Q. "Some Implementation Aspects of Sliding Window Least Squares Algorithms." IFAC Proceedings. Vol. 33, Issue 15, 2000, pp. 763–768.

Extended Capabilities

C/C++ Code Generation
Generate C and C++ code using Simulink® Coder™.

PLC Code Generation
Generate Structured Text code using Simulink® PLC Coder™.

See Also

|

Topics

Introduced in R2014a

System Identification Toolbox Documentation

Support