Standard deviation of Gauss-Markov process

[JoMeMi]

When investigating the Gauss-Markov process in Basilisk, I noticed the following lines:
GaussMarkov::GaussMarkov()
{
this->RNGSeed = 0x1badcad1;
std::normal_distribution<double>::param_type updatePair(0.0, 1.0/3.0);
this->rGen.seed((unsigned int)this->RNGSeed);
this->rNum.param(updatePair);
}

Why is it that the standard deviation of the normal distribution is 1/3? Is it always expected that the provided noiseMatrix represents 3 sigma, or is it some other reason? If I create a Gauss-Markov process on a three-dimensional state vector with a zero propMatrix, zero walkBounds and a diagonal noiseMatrix with 117e-6 in the diagonal, these are the results:

It definitely seems like 117e-6 is interpreted as 3 sigma, since most of the samples are below this threshold.

Secondly, I can see that in some of the sensor modules such as IMU and magnetometer, there is an additional scaling factor of 1.5. The documentation mentions this scaling factor, but says very little about why it is present. Are these two scaling factors connected it any way?

[AmolikaSoni]

Hi @JoMeMi, I have raised an issue here: #800 with a similar concern as yours. Do take a look.