Relating compute_paths CIRs to Scene's coverage_map in Sionna RT

[ifelse1984]

For Sionna RT, the CIRs and Coverage Maps are computed with different functions and classes. Is there a way to reconcile the two?

For instance, given a Tx (Array) + Rx pair, I can use compute path to get the CIR. Using a single Tx (Array), I can compute the coverage map, for which I can identify one of the cell centers that is closest to my Rx. Can I obtain the same path gain value (which seems time-invariant) on the coverage map from the CIR (which is time-variant)?

[jhoydis]

In short, the answer is no.

Coverage maps are computed as described here. They correspond to the normalized integral of the incoming energy across small surface patches. An approximation of this would be to compute the CIRs at several positions within a small region and compute their average energy. That means that you cannot perfectly recover the same path gain from a single CIR. However, they should be not too different if the cell size of the coverage map is not too large.

Why would you need to do this?

[ifelse1984]

Thank you for the details!

With regards to why:

1. If I were interested in only a few candidate positions/certain regions of the map, and I already have computed the CIRs, I was thinking of whether I have enough information to compute the coverage map (or an approximation of it) for that area without needing to re-run another ray-tracing routine.
2. More broadly, I am also hoping to understand better the relation between the two routines, and how they are related to each other despite being implemented by different methods (related to my follow-up question is below).

One thing that confuses me is that the channel coefficient h (in EM primer) is time varying, but h used in the coverage map computation (in the coverage map description) depends on position, but seems to be independent of time. Hence, I was wondering how do I reconcile the two, even in the limit where the cell-size is very small?