Issue with AoA in Sionna 1.0 when using higher max_depth

[tlstkddn999]

Hello,

I’m currently using Sionna 1.0 to run path simulations with the pathsolver, retrieving path data (including paths.vertices, paths.interaction, delay, channel_coefficient, AoD, and AoA) and filtering out only first-order reflections.

After that, I recreate the scenario in MATLAB to compare the reflection points obtained from Sionna with those calculated manually.

• Simulation workflow

1. Run a Sionna simulation to obtain the path data (paths.vertices, paths.interaction, delay, channel_coefficient, AoD, AoA, etc.).
2. the paths to extract only first-order reflections
3. Transfer the filtered path data into MATLAB. There, using the TX and RX positions along with the AoD/AoA from Sionna, I compute the reflection point myself and compare it with the Sionna-provided reflection point in paths.vertices.

• Results Observed

When max_depth = 1 (i.e., the simulation itself considers only one reflection):

1. The reflection points from Sionna (paths.vertices) match closely with my own MATLAB calculations.
2. The delay, channel_coefficient, AoD, and AoA values also align well between the two methods.

However, when max_depth > 1 (e.g., to include second-order or higher-order reflections) and I still filter down to “first-order reflections only” using the same approach:

1. The delay, channel_coefficient, and AoD remain consistent with the max_depth = 1 case.
2. The AoA often does not match: the path order can change, or the first-order reflection’s AoA may be missing or replaced. In particular, when max_depth is set to 5 or higher, the AoA for the first-order reflection sometimes disappears altogether.
   
   The plots above show only the first-order reflection components after filtering
   
   The image above shows all paths without filtering, and as can be seen, there are no lines from the RX passing through the reflection points — indicating that the AoA values might be incorrect.
   

I'm wondering if I might be misunderstanding something about how AoA is defined or computed in higher-order reflections in Sionna, or if there could be something wrong with how I’m setting up or interpreting the simulation.

I’d greatly appreciate any clarification or guidance. I hope my question and intention came through clearly.

Thank you!

[jhoydis]

Hi @tlstkddn999,

Thanks for reporting this issue. We will look into it and come back to you.

[faycalaa]

Hi @tlstkddn999,

Thank you for reporting this issue.
We were able to reproduce it, and it will be fixed in the next release.

Note: Related issue in the Sionna RT repository.

[faycalaa]

Version 1.0.2 should fix this issue.