Description of contribution in a few bullet points
Corrected the residual scaling.
I tested this in simulation and a few datasets and achieved good results. This might change some results or backend configurations, but I think it is more coherent with the scan matcher. In the end, the final uncertainty bounds attained depend on the tuning of the underlying probability model, which is heavily influenced by the covariance the scan matcher assigns.
Description of documentation updates required from your changes
Basic Info
Description of contribution in a few bullet points
I tested this in simulation and a few datasets and achieved good results. This might change some results or backend configurations, but I think it is more coherent with the scan matcher. In the end, the final uncertainty bounds attained depend on the tuning of the underlying probability model, which is heavily influenced by the covariance the scan matcher assigns.
Description of documentation updates required from your changes
None.
Future work that may be required in bullet points
None.