setMeasurementSizes

Sets the sizes of the measurement and measurement noise

Since R2024b

Syntax

setMeasurementSizes(filter,measurementSize,measurementNoiseSize)

Description

setMeasurementSizes(filter,measurementSize,measurementNoiseSize) sets the expected sizes of the measurement and measurement noise to the values in the measurementSize and measurementNoiseSize input arguments. When working with variable-sized measurements, use this function before passing measurements of new sizes to other object functions such as correct and distance.

Note

Only trackerGNN and trackerJPDA support variable-sized measurement, but not for the trackingKF filter.

Examples

Updating Measurement and Measurement Noise Size for Linear Kalman Filter

Create a linear Kalman filter that uses a 3D constant velocity motion model. Initialize the filter with a starting state vector indicating position and velocity.

KF = trackingKF(MotionModel="3D Constant Velocity" ...
    ,State=[100;0;0;0;0;0]);
predict(KF,1)

Predict the state of the filter after 1 second and use the distance function to lock the measurement model.

distance(KF,[100;0;0])

ans = 
3.5360

Display the current measurement model, measurement noise, and state covariance matrix.

disp(KF.MeasurementModel)

     1     0     0     0     0     0
     0     0     1     0     0     0
     0     0     0     0     1     0

disp(KF.MeasurementNoise)

     1     0     0
     0     1     0
     0     0     1

disp(KF.StateCovariance);

    2.2500    1.5000         0         0         0         0
    1.5000    2.0000         0         0         0         0
         0         0    2.2500    1.5000         0         0
         0         0    1.5000    2.0000         0         0
         0         0         0         0    2.2500    1.5000
         0         0         0         0    1.5000    2.0000

To update the measurement and measurement noise sizes for the filter, use the setMeasurementSizes function.

setMeasurementSizes(KF,2,2);

Update the measurement model of the Kalman filter to consider only the position in the x and z directions. Set the measurement noise to 3, which will scalar expand into a 2-by-2 matrix.

KF.MeasurementModel = [1 0 0 0 0 0; 0 0 1 0 0 0];
KF.MeasurementNoise = 3;

Display the updated measurement model and measurement noise.

disp(KF.MeasurementModel);

     1     0     0     0     0     0
     0     0     1     0     0     0

disp(KF.MeasurementNoise)

     3     0
     0     3

Use the distance function to lock the updated measurement model.

distance(KF,[100;0])

ans = 
3.3165

Correct the filter's state estimation using a new measurement vector with different size and display the updated state covariance matrix.

correct(KF,[100;0]);
disp(KF.StateCovariance);

    1.2857    0.8571         0         0         0         0
    0.8571    1.5714         0         0         0         0
         0         0    1.2857    0.8571         0         0
         0         0    0.8571    1.5714         0         0
         0         0         0         0    2.2500    1.5000
         0         0         0         0    1.5000    2.0000

Updating Measurement and Measurement Noise Size for Constant Velocity Extended Kalman Filter

Initialize the constant velocity Extended Kalman Filter (EKF) using an objectDetection object.

mp = struct('Frame', 'spherical', ...
    'HasElevation', false, 'HasVelocity', false);
detection = objectDetection(0,[0;100], ...
    'MeasurementParameters',mp);
filter = initcvekf(detection);

Predict the state of the filter after 1 second and use the distance function to lock the measurement function.

predict(filter, 1);
distance(filter,[0;100],{mp});

Display the filter's measurement noise and state covariance matrix.

disp(filter.MeasurementNoise);

     1     0
     0     1

disp(filter.StateCovariance);

   1.0e+03 *

    0.1013    0.1005         0         0         0         0
    0.1005    0.1010         0         0         0         0
         0         0    0.1033    0.1005         0         0
         0         0    0.1005    0.1010         0         0
         0         0         0         0    8.3371    0.1005
         0         0         0         0    0.1005    0.1010

To update the filter's measurement and noise sizes, use the setMeasurementSizes function.

setMeasurementSizes(filter,3,3);

Set the measurement noise to 2, which will scalar expand into a 3-by-3 matrix. Display the updated measurement noise for verification.

filter.MeasurementNoise = 2;
disp(filter.MeasurementNoise);

    2.0000         0         0
         0    2.0000         0
         0         0    2.0000

Use the distance function to lock the updated measurement function.

distance(filter,[100;0;0]) 

ans = 
18.3226

Correct the filter's state estimation using a new measurement vector with different size and display the updated state covariance matrix.

correct(filter,[100;0;0]);
disp(filter.StateCovariance);

    1.9613    1.9467         0         0         0         0
    1.9467    3.1768         0         0         0         0
         0         0    1.9620    1.9089         0         0
         0         0    1.9089    5.0777         0         0
         0         0         0         0    1.9995    0.0241
         0         0         0         0    0.0241   99.7888

Input Arguments

filter — Filter for object tracking

trackingKF object | trackingEKF object | trackingUKF object | trackingCKF object | trackingGSF object | trackingIMM object

Filter for object tracking, specified as one of these objects:

• trackingKF — Linear Kalman filter

• trackingEKF — Extended Kalman filter

• trackingUKF — Unscented Kalman filter

• trackingCKF — Cubature Kalman filter

• trackingGSF — Gaussian-sum filter

• trackingIMM — Interacting multiple model (IMM) filter

Note

When using setMeasurementSizes with trackingKF, the trackingKF filter must be initialized with the largest possible sizes for measurement and measurement noise.

measurementSize — Size of the measurement

positive real integer

Expected size of the measurement, specified as a positive real integer.

measurementNoiseSize — Size of the measurement noise

positive real integer

Expected size of the measurement noise, specified as a positive real integer.

Extended Capabilities

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

Version History

Introduced in R2024b