An error about non-integer nPotentials from inside of trackingJPDA function

When I use trackingJPDA in R2021a, I got an error about non-integer 'nPotentials'.

How can I avoid this error?

------------------ error ------------------

Error: repmat

Replication factors must be a row vector of integers or integer scalars.

Error: fusion.internal.assignment.feasibleJPDAEvents (Line 42)

FJE = repmat(omega,[1,1,nPotentials]);

Error: jpdaEvents (Line 70)

  [varargout{1:nargout}] = fusion.internal.assignment.feasibleJPDAEvents(varargin{:});

Error: trackerJPDA/calcEventAndProbabilities (Line 1881)

        FJE = obj.pEventGenerationFcn(validationMatrix);

Error: trackerJPDA/jpdaClusterUpdate (Line 1589)

        [FJE, FJEprob] = calcEventAndProbabilities(obj, clusterDetectionId, clusterTrackId, cMatrix);

Error: trackerJPDA/clusterAssignUpdate (Line 1447)

      [clusterInfo, isUnassignedDetection] = jpdaClusterUpdate(obj,preClusters,isUnassignedDetection);

Error: trackerJPDA/stepImpl (Line 1357)

        clusterAssignUpdate(obj,varargin{:});

Answer 1

[MathWorks Support Team]

This is a known issue in R2021a.

The problem comes from rounding off error in the factorial calculation on line 9 of numPotentialFeasibleEvents. There, we compute factorial(nM)/factorial(nM-nT+K)/factorial(k)

The workaround could be to add round() to that line like:

n1 = n1 + round(factorial(nT)/factorial(nT-k)*factorial(nM)/factorial(nM-nT+k)/factorial(k));