getTrackPositions

Returns updated track positions and position covariance matrix

Since R2021a

Syntax

positions = getTrackPositions(tracks,modelName)

positions = getTrackPositions(tracks,positionSelector)

[positions,positionCovariances] = getTrackPositions(___)

Description

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

positions = getTrackPositions(tracks,positionSelector) returns a matrix of track positions based on tracks and the position selector.

[positions,positionCovariances] = getTrackPositions(___) also returns the track position covariance matrices.

Examples

Find Position of 3-D Constant-Acceleration Object

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

tracker = radarTracker('FilterInitializationFcn',@initcaekf);

Update the tracker with a single detection and get the tracks output.

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

tracks = 
  objectTrack with properties:

                     TrackID: 1
                    BranchID: 0
                 SourceIndex: 0
                  UpdateTime: 0
                         Age: 1
                       State: [9×1 double]
             StateCovariance: [9×9 double]
             StateParameters: [1×1 struct]
               ObjectClassID: 3
    ObjectClassProbabilities: 1
                  TrackLogic: 'History'
             TrackLogicState: [1 0 0 0 0]
                 IsConfirmed: 1
                   IsCoasted: 0
              IsSelfReported: 1
            ObjectAttributes: [1×1 struct]

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

position1 = getTrackPositions(tracks,"constacc")

position1 = 1×3

    10   -20     4

Obtain the position vector from the track state using the position selector.

positionSelector = [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 0];
position2 = getTrackPositions(tracks,positionSelector)

position2 = 1×3

    10   -20     4

Find Position and Covariance of 3-D Constant-Velocity Object

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

tracker = radarTracker("FilterInitializationFcn",@initcvekf);

Update the tracker with a single detection and get the tracks output.

detection = objectDetection(0,[10;3;-7],"ObjectClassID",3);
tracks = tracker(detection,0)

tracks = 
  objectTrack with properties:

                     TrackID: 1
                    BranchID: 0
                 SourceIndex: 0
                  UpdateTime: 0
                         Age: 1
                       State: [6×1 double]
             StateCovariance: [6×6 double]
             StateParameters: [1×1 struct]
               ObjectClassID: 3
    ObjectClassProbabilities: 1
                  TrackLogic: 'History'
             TrackLogicState: [1 0 0 0 0]
                 IsConfirmed: 1
                   IsCoasted: 0
              IsSelfReported: 1
            ObjectAttributes: [1×1 struct]

Obtain the position vector and position covariance for that track using the model name.

[position1,positionCovariance1] = getTrackPositions(tracks,"constvel")

position1 = 1×3

    10     3    -7

positionCovariance1 = 3×3

     1     0     0
     0     1     0
     0     0     1

Obtain the position vector and position covariance for that track using the position selector.

positionSelector = [1 0 0 0 0 0; 0 0 1 0 0 0; 0 0 0 0 1 0];
[position2,positionCovariance2] = getTrackPositions(tracks,positionSelector)

position2 = 1×3

    10     3    -7

positionCovariance2 = 3×3

     1     0     0
     0     1     0
     0     0     1

Input Arguments

tracks — Object tracks

array of objectTrack objects | array of structures

Object tracks, specified as an array of objectTrack objects or an array of structures containing sufficient information to obtain the track position 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, positions and positionCovariances are returned based on the second argument (positionSelector or modelName) as follows.

Second input argument	positions	positionCovariances
postionSelector	positions is returned as a zeros(0,D) vector, where D is the row-dimension of the position selector. In this case, if the positonSelector is specified in single-precision, the positions output is in single-precision.	positionCovariances is returned as a zeros(D,D,0) matrix, where D is the row-dimension of the position selector. In this case, if the positonSelector is specified in single-precision, the positionCovariances output is in single-precision.
modelName	positions is returned as a zeros(0,3) vector.	positionCovariances is returned as a zeros(3,3,0) matrix.

modelName — Motion model name

"constvel" | "constacc" | "singer" | "constturn" | "ctrv"

Motion model name, specified as one of these options:

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

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

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

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

positionSelector — Position selection matrix

D-by-N real-valued matrix.

Position 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 positions from the state vector. Multiply the state vector by position selector matrix returns positions. The same selector is applied to all object tracks.

Output Arguments

positions — Positions of tracked objects

real-valued M-by-D matrix

Positions of tracked objects at last update time, returned as a real-valued M-by-D matrix. D represents the number of position elements. M represents the number of tracks.

positionCovariances — Position covariance matrices of tracked objects

real-valued D-by-D-M array

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

More About

Position Selector for 2-Dimensional Motion

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

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

Position Selector for 3-Dimensional Motion

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

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

Position Selector for 3-Dimensional Motion with Acceleration

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

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

Extended Capabilities

C/C++ Code Generation

Generate C and C++ code using MATLAB® Coder™.

• In code generation, if you use the modelName input, the tracks input must be specified as non-empty structures. Use the following if empty structures are unavoidable:

  if isempty(tracks)
     % Return the sizes and data types you desire
     positions = zeros(0,3,"single");
     positonCovariances = zeros(3,3,0,"single");
  else
     [positions,positionCovariances] = getTrackPositons(tracks,"constvel");
  end

Version History

Introduced in R2021a

R2022b: Obtain position and covariance from tracks using motion model name input

You can now obtain positions and associated covariances of tracks by specifying the motion model name as an input. For example,

[positions,covariances] = getTrackPositions(tracks,"constvel")

returns positions and position covariances in tracks based on the constant-velocity model in the constvel function.