Issue with unbiased residuals and KF tracks

[pbutti]

Seems like there is a mismatch between unbiased-biased residuals lists when filling the unbiased residuals histograms.
The code was originally written for ST + GBL assuming 3D hits (axial-stereo pair always present). In the case where one of the two sides is missing, there might be an issue with filling the plots.
My guess is that this loop is poorly defined for scatterOnly planes
https://github.com/JeffersonLab/hps-java/blob/master/tracking/src/main/java/org/hps/recon/tracking/gbl/SimpleGBLTrajAliDriver.java#L959
sensorMap will be filled for scatter only planes as well, but the nmeas will be 0. Hence will be skipped in filling the vector for biased residuals. I was expecting to skip the filling on the b_residual vector, but I might be wrong. We could test this in 13hits tracks.
@cbravo135 @tomeichlersmith @sarahgaiser