Revision history [back]

I would expect that map updates themselves do not slow down over time, as the computation of the voxels to "touch"/modify should take similar average computation time no matter where in the octomap the sensor is placed. Of course, if the average distance of range measurements increases over the course of your datasets, an increase in processing time is to be expected as the raytracing takes longer then. Of course, there are other factors at play that might have adverse effects, like allocation and placement of voxel information in memory.

What definitely takes longer the larger the map grows is iterating over all cells, for instance for visualization purposes. Are you possibly doing that somewhere in your code? If you are using the standard octomap_server , the publishAll() method is called after every map update. This method iterates over all cells, so it would explain a slow-down with increasing map size nicely.