Non euclidean distance in kernel home range estimation

Ciao,
I am trying to use an estimation of Kernel Home Range (i have 
adehabitate as reference package).

I would like to use a "distance" 
different from the euclidean one. I want to define a "friction" map so 
that the distance will be calculated as a cumulative cost path to 
travel between two points.
i can obtain this distance with the r.cost 
module in GRASS, but i would like to know if i can do the same with 
some R package...
Any idea or any suggestion about this issue?
Thanks a 
lot

Ferdinando