Main Content

reset

Reset online parameter estimation System object

Description

example

reset(obj) resets the states of a locked online parameter estimation System object™, obj, to initial values and leaves the object locked. The states of the object are the estimated parameters and parameter covariance. Use reset if you are not satisfied with the estimation or if your system changes modes.

Examples

collapse all

Create a System object for online estimation of an Output-Error model.

obj = recursiveOE('InitialB',[0 0.5],'InitialF',[1 0.8],...
    'InitialParameterCovariance',0.1);

Load the estimation data. For this example, use a static data set for illustration.

load iddata1 z1;
output = z1.y;
input = z1.u;

Estimate model parameters online using step.

for i = 1:numel(input)
  [B,F,EstimatedOutput] = step(obj,output(i),input(i));
end

View the object properties.

obj
obj = 
  recursiveOE with properties:

                             B: [0 2.0014]
                             F: [1 -0.7639]
                      InitialB: [0 0.5000]
                      InitialF: [1 0.8000]
           ParameterCovariance: [2x2 double]
    InitialParameterCovariance: [2x2 double]
              EstimationMethod: 'ForgettingFactor'
              ForgettingFactor: 1
              EnableAdaptation: true
                       History: 'Infinite'
               InputProcessing: 'Sample-based'
                      DataType: 'double'

Reset the System object.

reset(obj)

The estimated parameters, B and F, and parameter covariance, ParameterCovariance are reset to the initial values.

obj
obj = 
  recursiveOE with properties:

                             B: [0 0.5000]
                             F: [1 0.8000]
                      InitialB: [0 0.5000]
                      InitialF: [1 0.8000]
           ParameterCovariance: [2x2 double]
    InitialParameterCovariance: [2x2 double]
              EstimationMethod: 'ForgettingFactor'
              ForgettingFactor: 1
              EnableAdaptation: true
                       History: 'Infinite'
               InputProcessing: 'Sample-based'
                      DataType: 'double'

Input Arguments

collapse all

System object for online parameter estimation, created using one of the following commands:

Online Estimation System ObjectEstimated Parameters
recursiveAR

A — Reset to InitialA

recursiveARMA

A — Reset to InitialA

C — Reset to InitialC

recursiveARX

A — Reset to InitialA

B — Reset to InitialB

recursiveARMAX

A — Reset to InitialA

B — Reset to InitialB

C — Reset to InitialC

recursiveOE

B — Reset to InitialB

F — Reset to InitialF

recursiveBJ

B — Reset to InitialB

C — Reset to InitialC

D — Reset to InitialD

F — Reset to InitialF

recursiveLS

Parameters — Reset to InitialParameters

When EstimationMethod property of obj is 'ForgettingFactor' or 'KalmanFilter', the ParameterCovariance property of obj is reset to the value of InitialParameterCovariance.

Introduced in R2015b