Scattering in RL-GO

As I was using the RL-GO solver in a radar application, I noticed that it casts rays from the source that do not end at the "receiver" (the near field sampler). Instead, the end of a ray is denoted with an "S," which the documentation describes as an "observation point." Here is a picture of one ray:

My source is the dipole in free space above the dielectric cuboid. My near field sampler is co-located with my source, and I have checked the setting to "calculate only the scattered part of the field."

There is no description of this "observation point" concept in the documentation. I am curious how a ray like the one above contributes to the final electric field observed by the receiver. Is RL-GO considering scattering from each observation point? If so, how does the solver calculate the portion of energy that returns to the receiver?