The distance to an obstacle computation (if using full footprint checking) is lower bound by the inscribed radius.
The inflation layer therefore doesn't support distance computations for a full footprint in situations where it would be interesting (e.g. driving along walls).
For example, irrespective of the orientation of the red footprint we would get the same distance result (=inscribed radius, in this case many times larger than the actual distance) and even though we might get arbitrarily close to an obstacle, collision_margin_distancewould never be reached and the assumed distance would be the inscribed radius (in this case 0.6m).
My point being: I fail to see the use of the footprint distance checking. Imho would be of great help, but only given that we computed the actual distances.
Do I have a knot in my head, or does it make sense what I am saying here. Does it make sense to try to calculate the actual distances efficiently?
The distance to an obstacle computation (if using full footprint checking) is lower bound by the inscribed radius.
The inflation layer therefore doesn't support distance computations for a full footprint in situations where it would be interesting (e.g. driving along walls).
For example, irrespective of the orientation of the red footprint we would get the same distance result (=inscribed radius, in this case many times larger than the actual distance) and even though we might get arbitrarily close to an obstacle,
collision_margin_distancewould never be reached and the assumed distance would be the inscribed radius (in this case 0.6m).My point being: I fail to see the use of the footprint distance checking. Imho would be of great help, but only given that we computed the actual distances.
Do I have a knot in my head, or does it make sense what I am saying here. Does it make sense to try to calculate the actual distances efficiently?