Closed xyaoab closed 11 months ago
Because unsigned chars are the underlying data in the costmap, not doubles. Its 1/8 the size allowing it to scale better. I won't say that that'll always be the case, but it was a solid engineering decision back with the resources we had in the 2000s when Willow was working on that package. Its on my agenda in 2024 to enable more flexibility here, but today, unsigned chars are your options. 0-100 is also what we think about in the literature for probability fields, which 0-255 handles quite nicely.
Hi team,
Is there any specific reason that
unsigned char
instead ofdouble
is supported in cost values in nav2_costmap_2d /costmap_2d.hpp?Feature request
Feature description
We are using distance to the nearest occupied cell to represent the cost value in the costmap.
double
is expected to have higher precision.