Inconsistent sensor noise handling

[sdunlap-afit]

Hello, I saw #862 was merged and took a look at the sensor implementations. I am noticing some issues/inconsistencies in the latest develop branch. I have only tested a couple of these, the rest were from code review. Some of these may be intentional, but I wanted to point them out.

1. magnetometer.cpp: There is no call to noiseModel.setPropMatrix().
2. coarseSunSensor.cpp:
   • Line 138 still has the 1.5x multiplier for the noise model, but line 159 (fault model) does not. Also, the other sensors no longer have this multiplier.
   • There is no way for the user to set the propagation matrix for the noise model.
3. simpleVoltEstimator.h: Line 56 AMatrix is private, so there's no way for the user to set the propagation matrix for the noise model.
4. starTracker.h: Line 72 AMatrix is also private.
5. imuSensor.cpp: Having to set AMatrix after initializing the simulation is pretty inconvenient. Would it be possible to set it in the constructor instead of in Reset?

Given the above, would it make sense for each sensor to initialize the GaussMarkov object with reasonable defaults in the constructor, and then simply expose (mark public) the GaussMarkov object to the user for further customization? It seems like this would make it easier to keep the various sensors consistent and would make it easier for the user to customize the noise model.

[Natsoulas]

Good catch on coarseSunSensor.cpp! I totally missed that. I'll have to get back to you about your other questions about sensor implementations. Definitely might be good to reconcile them.

[Natsoulas]

Fixed missed sun sensor 1.5x multiplier removal. Will have an update soon on your other items.

[sdunlap-afit]

Excellent, thanks for the quick response!

[Natsoulas]

@sdunlap-afit I will be refactoring sensors' implementation of noise and addressing a good chunk of your items. Thanks again for pointing these inconsistencies out. Do add comments here if anything else relevant comes up.

[sdunlap-afit]

@Natsoulas Thanks for working on this. I was looking at scenarioGaussMarkovRandomWalk.py and noticed that imuSensor1.AMatrixGyro is set before scSim.InitializeSimulation(), so Reset() is reverting it to I. Also, do negative values make sense for a random walk?