While compiling BFL, I noticed this warning from the compiler
[...]
g++ -DHAVE_CONFIG_H -I. -I../../../src/filter -I../.. -D
__RNGWRAPPER_LTI__ -I/usr/include -g -O2 -I/usr//include/matrixwrapper
-I/usr/include/ltilib/ -MT EKparticlefilter.lo -MD -MP -MF
.deps/EKparticlefilter.Tpo -c ../../../src/filter/EKparticlefilter.cpp
-o EKparticlefilter.o
../../../src/filter/iteratedextendedkalmanfilter.cpp: In member
function 'virtual void
BFL::IteratedExtendedKalmanFilter::MeasUpdate(BFL::MeasurementModel
const MatrixWrapper::ColumnVector&)':
../../../src/filter/iteratedextendedkalmanfilter.cpp:91: warning:
passing 'double' for argument 1 to
'MatrixWrapper::ColumnVector::ColumnVector(int)'
InnovationCheck::check() expects a ColumnVector, and a double is
passed. It seems to me (I didn't check very carefully though) some
work
>From IEKF.cpp
===
ColumnVector innovation = (x_i - x_i_prev);
norm_inno = innovation.transpose() * innovation;
if (innovationChecker != NULL)
test_innovation = innovationChecker->check(norm_inno) ; //test
if the innovation is not too small
===
===
In InnovationCheck.cpp
bool
InnovationCheck::check(ColumnVector innovation)
{
// basic implementation which checks if the norm of the innovation is
// higher than the minimum innovation specified
return (innovation.transpose() * innovation >= min_innovation);
}
====
Is this a bug or did I overlook something?
Klaas
ps. The InnovationCheck::check function is not API documented.
_______________________________________________
I hereby promise not to top-post on the
BFL mailing list
BFL [..] ...
http://lists.mech.kuleuven.be/mailman/listinfo/bfl
IEKF InnovationCheck
> Is this a bug or did I overlook something?
>
> Klaas
>
> ps. The InnovationCheck::check function is not API documented.
Solved (also the api documentation) in revision 8251.
_________________________________________________________
svn commit
Sending ChangeLog
Sending src/filter/innovationCheck.h
Sending src/filter/iteratedextendedkalmanfilter.cpp
Transmitting file data ...
Committed revision 8251.
_________________________________________________________
Tinne
_______________________________________________
I hereby promise not to top-post on the
BFL mailing list
BFL [..] ...
http://lists.mech.kuleuven.be/mailman/listinfo/bfl
Disclaimer: http://www.kuleuven.be/cwis/email_disclaimer.htm