# getTrackVelocities

Obtain updated track velocities and velocity covariance matrix

Since R2018b

## Syntax

``positions = getTrackVelocities(tracks,modelName)``
``velocities = getTrackVelocities(tracks,velocitySelector)``
``````[velocities,velocityCovariances] = getTrackVelocities(tracks,velocitySelector)``````

## Description

example

````positions = getTrackVelocities(tracks,modelName)` returns a matrix of track velocities based on tracks and the model name.```

example

````velocities = getTrackVelocities(tracks,velocitySelector)` returns a matrix of track velocities based on tracks and the velocity selector.```

example

``````[velocities,velocityCovariances] = getTrackVelocities(tracks,velocitySelector)``` also returns the track velocity covariance matrices.```

## Examples

collapse all

Create an extended Kalman filter tracker for 3-D constant-acceleration motion.

`tracker = trackerGNN("FilterInitializationFcn",@initcaekf);`

Initialize the tracker with one detection.

```detection = objectDetection(0,[10;-20;4],"ObjectClassID",3); tracks = tracker(detection,0);```

Add a second detection at a later time and at a different position.

```detection = objectDetection(0.1,[10.3;-20.2;4],"ObjectClassID",3); tracks = tracker(detection,0.2);```

Obtain the velocity vector from the track state using the model name.

`velocity1 = getTrackVelocities(tracks,"constacc")`
```velocity1 = 1×3 1.0093 -0.6728 0 ```

Obtain the velocity vector from the track state using the positin selector.

```velocitySelector = [0 1 0 0 0 0 0 0 0; 0 0 0 0 1 0 0 0 0; 0 0 0 0 0 0 0 1 0]; velocity2 = getTrackVelocities(tracks,velocitySelector)```
```velocity2 = 1×3 1.0093 -0.6728 0 ```

Create an extended Kalman filter tracker for 3-D constant-acceleration motion.

`tracker = trackerGNN("FilterInitializationFcn",@initcaekf);`

Initialize the tracker with one detection.

```detection = objectDetection(0,[10;-20;4],"ObjectClassID",3); tracks = tracker(detection,0);```

Add a second detection at a later time and at a different position.

```detection = objectDetection(0.1,[10.3;-20.2;4.3],"ObjectClassID",3); tracks = tracker(detection,0.2);```

Obtain the velocity vector and covariance from the track state using the model name.

`[velocity1,velocityCovariance1] = getTrackVelocities(tracks,"constacc")`
```velocity1 = 1×3 1.0093 -0.6728 1.0093 ```
```velocityCovariance1 = 3×3 70.0685 0 0 0 70.0685 0 0 0 70.0685 ```

Obtain the velocity vector and covariance from the track state using the velocity selector.

```velocitySelector = [0 1 0 0 0 0 0 0 0; 0 0 0 0 1 0 0 0 0; 0 0 0 0 0 0 0 1 0]; [velocity2,velocityCovariance2] = getTrackVelocities(tracks,velocitySelector)```
```velocity2 = 1×3 1.0093 -0.6728 1.0093 ```
```velocityCovariance2 = 3×3 70.0685 0 0 0 70.0685 0 0 0 70.0685 ```

## Input Arguments

collapse all

Object tracks, specified as an array of `objectTrack` objects or an array of structures containing sufficient information to obtain the track velocity information. At a minimum, these structures must contain a `State` column vector field and a positive-definite `StateCovariance` matrix field. For a sample track structure, see `toStruct`.

Note

If you specify `tracks` as an empty `objectTrack` object, an empty cell, or an empty track structure, `velocities` and `velocityCovariances` are returned based on the second argument (`velocitySelector` or `modelName`) as follows.

Second input argument`velocities``velocityCovariances`
`velocitySelector`

`velocities` is returned as a `zeros(0,D)` vector, where `D` is the row-dimension of the velocity selector. In this case, if the `velocitySelector` is specified in single-precision, the `positions` output is in single-precision.

`velocityCovariances` is returned as a `zeros(D,D,0)` matrix, where `D` is the row-dimension of the velocity selector. In this case, if the `velocitySelector` is specified in single-precision, the `velocityCovariances` output is in single-precision.

`modelName`

`velocities` is returned as a `zeros(0,3)` vector.

`velocityCovariances` is returned as a `zeros(3,3,0)` matrix.

Motion model name, specified as one of these options:

• `"constvel"` — The function obtains the velocity states based on the state definition in the `constvel` function.

• `"constacc"` — The function obtains the velocity states based on the state definition in the `constacc` function.

• `"constturn"` — The function obtains the velocity states based on the state definition in the `constturn` function.

• `"singer"` — The function obtains the velocity states based on the state definition in the `singer` function. The use of `singer` model requires the Sensor Fusion and Tracking Toolbox™.

Velocity selector, specified as a D-by-N real-valued matrix of ones and zeros. D is the number of dimensions of the tracker. N is the size of the state vector. Using this matrix, the function extracts track velocities from the state vector. Multiply the state vector by velocity selector matrix returns velocities. The same selector is applied to all object tracks.

## Output Arguments

collapse all

Velocities of tracked objects at last update time, returned as a 1-by-D vector or a real-valued M-by-D matrix. D represents the number of velocity elements. M represents the number of tracks.

Velocity covariance matrices of tracked objects, returned as a real-valued D-by-D-matrix or a real-valued D-by-D-by-M array. D represents the number of velocity elements. M represents the number of tracks. Each D-by-D submatrix is a velocity covariance matrix for a track.

collapse all

### Velocity Selector for 2-Dimensional Motion

Show the velocity selection matrix for two-dimensional motion when the state consists of the position and velocity.

`$\left[\begin{array}{cccc}0& 1& 0& 0\\ 0& 0& 0& 1\end{array}\right]$`

### Velocity Selector for 3-Dimensional Motion

Show the velocity selection matrix for three-dimensional motion when the state consists of the position and velocity.

`$\left[\begin{array}{cccccc}0& 1& 0& 0& 0& 0\\ 0& 0& 0& 1& 0& 0\\ 0& 0& 0& 0& 0& 1\end{array}\right]$`

### Velocity Selector for 3-Dimensional Motion with Acceleration

Show the velocity selection matrix for three-dimensional motion when the state consists of the position, velocity, and acceleration.

`$\left[\begin{array}{ccccccccc}0& 1& 0& 0& 0& 0& 0& 0& 0\\ 0& 0& 0& 0& 1& 0& 0& 0& 0\\ 0& 0& 0& 0& 0& 0& 0& 1& 0\end{array}\right]$`

## Version History

Introduced in R2018b

expand all