delay_spread_from_rays#

sionna.phy.channel.tr38901.delay_spread_from_rays(rays, lsp, scenario, serving: torch.Tensor | None = None, include_los_component: bool = True, include_subclusters: bool = True) → torch.Tensor[source]#

Computes delay spread from 3GPP TR 38.901 [TR38901V1920] rays.

This helper applies the LoS K-factor weighting from Eq. (7.5-30) and the optional sub-cluster expansion from Eq. (7.5-26) and Table 7.5-5 of [TR38901V1920] before evaluating rms_delay_spread().

Parameters:

  • rays – Rays. Rays for all BS-UT links.

  • lsp – LSP. Large-scale parameters associated with rays.

  • scenario – SystemLevelScenario. Scenario used to generate rays and lsp.

  • serving (torch.Tensor | None) – Serving BS indices with shape [batch size, num_ut] or [batch size*num_ut]. If None, delay spread is computed for all BS-UT links.

  • include_los_component (bool) – If True, include the deterministic LoS component according to the K-factor.

  • include_subclusters (bool) – If True, split the two strongest clusters according to the subcluster delay offsets from [TR38901V1920].

Outputs:

ds – Delay spread [s]. Shape is [batch size, num_ut] if serving is provided and [batch size, num_bs, num_ut] otherwise.