This PR fixes the issue that samplePoseGaussian() is currently drawing positions from the uniform distribution instead of the Gaussian distribution.
The Problem
Currently samplePoseGaussian() samples positions uniformly between -pos_variances and +pos_variances.
Expected Behavior
samplePoseGaussian() should sample from the normal distribution: N(0, pos_variances).
The Fix
This PR fixes the issue by making samplePoseGaussian() call samplePositionGaussian() instead of samplePositionUniform().
Other changes
The input parameter to samplePoseGaussian() has also been renamed from pos_variances to pos_stddev to reflect that the input should be the standard deviation, not the variance (as reflected in the documentation of samplePositionGaussian()).
Fixed a discrepency between the input parameter name in the function signature of sampleOrientationUniform() in tf.h and tf.cpp.
This PR fixes the issue that
samplePoseGaussian()
is currently drawing positions from the uniform distribution instead of the Gaussian distribution.The Problem
Currently
samplePoseGaussian()
samples positions uniformly between-pos_variances
and+pos_variances
.Expected Behavior
samplePoseGaussian()
should sample from the normal distribution:N(0, pos_variances)
.The Fix
This PR fixes the issue by making
samplePoseGaussian()
callsamplePositionGaussian()
instead ofsamplePositionUniform()
.Other changes
The input parameter to
samplePoseGaussian()
has also been renamed frompos_variances
topos_stddev
to reflect that the input should be the standard deviation, not the variance (as reflected in the documentation ofsamplePositionGaussian()
).Fixed a discrepency between the input parameter name in the function signature of
sampleOrientationUniform()
intf.h
andtf.cpp
.