pith. sign in

arxiv: 2604.09243 · v1 · submitted 2026-04-10 · 💻 cs.CE

BVH-Accelerated Ray Tracing for High-Frequency Electromagnetic Backscattering

Marco Pasquale , Andong Hu , Luca Pennati , Ivy Peng , Stefano Markidis This is my paper

Pith reviewed 2026-05-10 16:30 UTC · model grok-4.3

classification 💻 cs.CE
keywords shooting and bouncing raysbounding volume hierarchyphysical opticsradar cross sectionhigh-frequency electromagneticselectromagnetic backscatteringGPU acceleration
0
0 comments X

The pith

A bounding volume hierarchy accelerates shooting-and-bouncing-rays computation of high-frequency electromagnetic backscattering from metallic objects.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

Full-wave electromagnetic solvers grow prohibitively expensive as frequency rises and objects become electrically large. The paper develops a shooting-and-bouncing-rays method that traces geometrical-optics rays for multiple reflections and then evaluates a physical-optics surface integral over the resulting ray tubes. A bounding-volume hierarchy prunes the intersection search, the workload is cast as a trace-integrate pipeline, and an incident-ray sampling rule prevents phase aliasing in the discretized integral. The implementation runs on both NVIDIA and AMD GPUs and is distributed with MPI; results are checked against exact Mie solutions for a conducting sphere and applied to an aircraft shape for monostatic radar-cross-section prediction.

Core claim

The authors present a shooting-and-bouncing-rays method that couples multi-reflection geometrical-optics ray transport with a physical-optics surface integral discretized over ray tubes. A bounding-volume hierarchy reduces the ray-surface intersection search space, the computation is organized as a trace-integrate pipeline, and numerical accuracy is maintained by an incident-ray sampling rule that mitigates phase aliasing. The algorithm is accelerated on NVIDIA and AMD GPUs and parallelized with MPI, with validation against analytical Mie solutions for a perfectly electrically conducting sphere and demonstration on a complex aircraft geometry for monostatic radar-cross-section prediction.

What carries the argument

Bounding-volume hierarchy (BVH) that accelerates ray-surface intersection queries inside a trace-integrate pipeline for shooting-and-bouncing-rays transport combined with physical-optics integration.

If this is right

  • High-frequency backscattering from electrically large metallic objects becomes computationally feasible where full-wave methods are intractable.
  • Monostatic radar-cross-section predictions for complex geometries such as aircraft can be obtained with controlled accuracy.
  • The same trace-integrate structure runs efficiently on both NVIDIA and AMD GPUs and scales across MPI processes.
  • Validation on analytic sphere solutions provides a concrete benchmark that any later extension must reproduce.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The same BVH-accelerated pipeline could be reused for bistatic or time-domain scattering problems by changing only the observation direction and integration kernel.
  • If the sampling rule proves robust, the method offers a practical route to embedding high-frequency scattering inside larger system-level simulations that already use ray tracing.
  • Parallel performance on mixed NVIDIA/AMD hardware suggests the approach could be adopted in existing GPU-accelerated ray-tracing libraries without major architectural changes.

Load-bearing premise

The chosen incident-ray sampling rule sufficiently mitigates phase aliasing in the discretized physical-optics integration to keep numerical accuracy acceptable.

What would settle it

Compute the monostatic radar cross-section of a perfectly conducting sphere at several increasing frequencies and compare the result with the exact Mie-series solution; a discrepancy that grows with frequency or exceeds expected discretization error would falsify the accuracy claim.

Figures

Figures reproduced from arXiv: 2604.09243 by Andong Hu, Ivy Peng, Luca Pennati, Marco Pasquale, Stefano Markidis.

Figure 1
Figure 1. Figure 1: Summary of the SBR pipeline. Rays are launched from the orthographic grid, the BVH ac￾celerates hit detection and traversal by reducing the search space. The scattered field contributions are then accumulated to compute the RCS of the object [PITH_FULL_IMAGE:figures/full_fig_p005_1.png] view at source ↗
Figure 3
Figure 3. Figure 3: Visualizing a BVH constructed using the Binned SAH method. Each bounding AABB is created [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Frequency scan of monostatic RCS for a triangulated PEC sphere, averaged across incident [PITH_FULL_IMAGE:figures/full_fig_p008_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Radar cross-section (RCS) calculation for an A380 aircraft. The simulation uses 500 samples in [PITH_FULL_IMAGE:figures/full_fig_p009_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Aircraft geometry and radar cross section representation. (a) Simulated A380 aircraft model with [PITH_FULL_IMAGE:figures/full_fig_p010_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Execution trace from NVIDIA Nsight Systems on 2 MPI processes each with one NVIDIA A100. [PITH_FULL_IMAGE:figures/full_fig_p011_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: Scaling performance on LUMI, using FP32. On the [PITH_FULL_IMAGE:figures/full_fig_p012_8.png] view at source ↗
read the original abstract

As computational complexity in electromagnetics increases with frequency, full-wave solvers become computationally infeasible for electrically large problems. To address this limitation, we present a shooting and bouncing rays (SBR) method for efficiently modeling electromagnetic backscattering of metallic objects in the high-frequency regime. The method couples multi-reflection geometrical-optics ray transport with a physical optics surface integral discretized over ray tubes. To reduce the massive ray-surface intersection search space, we use a bounding volume hierarchy (BVH) and organize the computation as a trace-integrate pipeline. The ray tracing generates hit data, and the physical optics integral is evaluated over valid intersections only. Numerical accuracy is controlled through an incident-ray sampling rule that mitigates phase aliasing in the discretized physical optics integration. The method is accelerated on NVIDIA and AMD GPUs and parallelized with MPI. We validate against analytical Mie solutions for a perfectly electrically conducting (PEC) sphere and demonstrate applicability to a complex aircraft geometry for monostatic radar cross-section prediction.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 1 minor

Summary. The paper presents a BVH-accelerated shooting-and-bouncing-rays (SBR) method that couples geometrical-optics ray transport with a physical-optics surface integral discretized over ray tubes for high-frequency electromagnetic backscattering from metallic objects. The approach uses a trace-integrate pipeline, is parallelized on NVIDIA/AMD GPUs and with MPI, controls accuracy via an incident-ray sampling rule intended to mitigate phase aliasing, validates against analytical Mie solutions for a PEC sphere, and demonstrates monostatic RCS prediction on an aircraft geometry.

Significance. If the accuracy claims hold, the work supplies a practical, scalable implementation of standard high-frequency approximations for electrically large scatterers where full-wave solvers are infeasible. The BVH acceleration, GPU/MPI parallelization, and demonstration on complex aircraft geometry constitute engineering strengths that could support radar-cross-section applications.

major comments (2)
  1. [Abstract and Methods (incident-ray sampling rule)] The incident-ray sampling rule is presented only as a heuristic density choice in the abstract and methods; no derivation relating sampling interval to local phase gradient, no truncation-error estimate for the discretized physical-optics integral, and no verification that the rule remains sufficient once multiple bounces and shadowing are present. This is load-bearing for the central accuracy claim.
  2. [Validation] The validation section reports agreement with Mie solutions for a PEC sphere but supplies no quantitative error metrics, convergence data with respect to sampling density, or implementation details (e.g., ray-tube discretization parameters), leaving the support for numerical-accuracy claims difficult to evaluate.
minor comments (1)
  1. [Abstract] The abstract and introduction could more explicitly state the frequency range or electrical size of the target objects to clarify the high-frequency regime in which the method is intended to operate.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive and detailed review of our manuscript on the BVH-accelerated SBR method for high-frequency electromagnetic backscattering. The comments highlight important areas for improving the rigor of our accuracy claims, and we address each major comment below with plans for revision.

read point-by-point responses

  1. Referee: [Abstract and Methods (incident-ray sampling rule)] The incident-ray sampling rule is presented only as a heuristic density choice in the abstract and methods; no derivation relating sampling interval to local phase gradient, no truncation-error estimate for the discretized physical-optics integral, and no verification that the rule remains sufficient once multiple bounces and shadowing are present. This is load-bearing for the central accuracy claim.

    Authors: We agree that the sampling rule requires more rigorous justification in the manuscript. The rule is motivated by the need to sample the incident field at a density sufficient to capture phase variations without aliasing in the physical-optics integral over ray tubes, but the current text presents it concisely without explicit derivation or error bounds. In the revised manuscript we will add a dedicated subsection deriving the sampling interval from the local phase gradient of the incident wave (based on a first-order Taylor expansion of the phase term), provide a truncation-error estimate for the discretized surface integral, and include numerical verification that the chosen density remains adequate under multiple bounces and shadowing by reporting results on the aircraft geometry with varying bounce counts. revision: yes

  2. Referee: [Validation] The validation section reports agreement with Mie solutions for a PEC sphere but supplies no quantitative error metrics, convergence data with respect to sampling density, or implementation details (e.g., ray-tube discretization parameters), leaving the support for numerical-accuracy claims difficult to evaluate.

    Authors: We acknowledge that the validation currently relies on qualitative visual agreement rather than quantitative metrics. In the revised version we will augment the validation section with root-mean-square error values between the computed monostatic RCS and the analytical Mie series for the PEC sphere at multiple frequencies, convergence plots of error versus sampling density, and explicit implementation details such as the number of quadrature points per ray tube, the precise sampling interval in wavelengths, and the ray-tube width parameter. These additions will directly support the accuracy claims and allow independent evaluation of the method. revision: yes

Circularity Check

0 steps flagged

No circularity: derivation is a composition of standard methods with external validation

full rationale

The paper composes standard SBR (geometrical-optics ray tracing + physical-optics surface integral) with BVH acceleration and MPI/GPU parallelization. Accuracy control is asserted via an incident-ray sampling rule whose sufficiency is shown by direct comparison to the analytical Mie solution for a PEC sphere; no equation, parameter, or claim is shown to reduce to a fitted input, self-definition, or self-citation chain. The derivation chain therefore remains independent of its own outputs and is externally falsifiable.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

Based on abstract only; the approach rests on standard high-frequency EM approximations whose validity is assumed rather than re-derived.

axioms (1)
  • domain assumption High-frequency regime permits separation into geometrical-optics ray transport and physical-optics surface integral evaluation.
    Invoked throughout the method description as the basis for the SBR-PO coupling.

pith-pipeline@v0.9.0 · 5476 in / 1344 out tokens · 82975 ms · 2026-05-10T16:30:27.362612+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

24 extracted references · 24 canonical work pages

  1. [1]

    Computation of Antenna Transfer Functions with a Bidirectional Ray-Tracing Algorithm Utilizing Antenna Reciprocity

    M. M. Taygur, I. O. Sukharevsky, and T. F. Eibert. “Computation of Antenna Transfer Functions with a Bidirectional Ray-Tracing Algorithm Utilizing Antenna Reciprocity”. en. In:2018 2nd URSI Atlantic Radio Science Meeting (AT-RASC). Gran Canaria: IEEE, May 2018, pp. 1–4.doi:10.23919/URSI- AT-RASC.2018.8471651.url:https://ieeexplore.ieee.org/document/8471651/. 12

    work page doi:10.23919/ursi- 2018

  2. [2]

    A Bidirectional Ray-Tracing Method for Antenna Coupling Evaluation Based on the Reciprocity Theorem

    M. M. Taygur, I. O. Sukharevsky, and T. F. Eibert. “A Bidirectional Ray-Tracing Method for Antenna Coupling Evaluation Based on the Reciprocity Theorem”. en. In:IEEE Transactions on Antennas and Propagation66.12 (Dec. 2018), pp. 6654–6664.doi:10 . 1109 / TAP . 2018 . 2876680.url: https://ieeexplore.ieee.org/document/8496863/

    work page arXiv 2018

  3. [3]

    IEEE Access 3, 1089–1100 (2015) https://doi.org/10.1109/ACCESS.2015.2453991

    Z. Yun and M. F. Iskander. “Ray Tracing for Radio Propagation Modeling: Principles and Applica- tions”. en. In:IEEE Access3 (2015), pp. 1089–1100.doi:10.1109/ACCESS.2015.2453991.url: http://ieeexplore.ieee.org/document/7152831/

    work page doi:10.1109/access.2015.2453991.url: 2015

  4. [4]

    GPU accelerated ray launching for high-fidelity virtual test drives of V ANET applications

    M. Schiller et al. “GPU accelerated ray launching for high-fidelity virtual test drives of V ANET applications”. en. In:2015 International Conference on High Performance Computing & Simulation (HPCS). Amsterdam, Netherlands: IEEE, July 2015, pp. 262–268.doi:10.1109/HPCSim.2015. 7237048.url:http://ieeexplore.ieee.org/document/7237048/

    work page doi:10.1109/hpcsim.2015 2015

  5. [5]

    Parallel GPU Optimization of the Shooting and Bouncing Ray Tracing Methodology for Propagation Modeling

    S. Kasdorf et al. “Parallel GPU Optimization of the Shooting and Bouncing Ray Tracing Methodology for Propagation Modeling”. In:IEEE Transactions on Antennas and Propagation72.1 (2024), pp. 174– 182.doi:10.1109/TAP.2024.3350896

    work page doi:10.1109/tap.2024.3350896 2024

  6. [6]

    Accelerated Ray Launching Method for Efficient Field Coverage Studies in Wide Urban Areas

    J. G ´omez et al. “Accelerated Ray Launching Method for Efficient Field Coverage Studies in Wide Urban Areas”. en. In:Sensors23.14 (July 2023), p. 6412.doi:10.3390/s23146412.url:https: //www.mdpi.com/1424-8220/23/14/6412

    work page doi:10.3390/s23146412.url:https: 2023

  7. [7]

    Combination of a Full-Wave Method and Ray Tracing for Radiation Pattern Simulations of Antennas on Vehicle Roofs

    T. J. Purcell et al. “Combination of a Full-Wave Method and Ray Tracing for Radiation Pattern Simulations of Antennas on Vehicle Roofs”. en. In:ACM Transactions on Graphics21.3 (July 2002), pp. 703–712.doi:10 . 1145 / 566654 . 566640.url:https : / / dl . acm . org / doi / 10 . 1145 / 566654.566640

    work page arXiv 2002

  8. [8]

    Exploring the key technologies and applications of 6G wireless communication network

    P. Li, J. Fan, and J. Wu. “Exploring the key technologies and applications of 6G wireless communication network”. en. In:iScience28.5 (May 2025), p. 112281.doi:10.1016/j.isci.2025.112281.url: https://linkinghub.elsevier.com/retrieve/pii/S2589004225005425

    work page doi:10.1016/j.isci.2025.112281.url: 2025

  9. [10]

    Shooting and bouncing rays: Calculating RCS of an arbitrary cavity

    H. Ling, R. Chou, and S. Lee. “Shooting and bouncing rays: Calculating RCS of an arbitrary cavity”. In: 1986 Antennas and Propagation Society International Symposium. Vol. 24. June 1986, pp. 293–296. doi:10.1109/APS.1986.1149823.url:https://ieeexplore.ieee.org/document/1149823

    work page doi:10.1109/aps.1986.1149823.url:https://ieeexplore.ieee.org/document/1149823 1986

  10. [11]

    Computational Electromagnetics: Software Development and High Frequency Modelling of Surface Currents on Perfect Conductors

    S. Sefi. “Computational Electromagnetics: Software Development and High Frequency Modelling of Surface Currents on Perfect Conductors”. DiV A portal. PhD thesis. Stockholm: KTH Royal Institute of Technology, 2005

    work page 2005

  11. [12]

    Doyle, Michael Guthe, and Jiří Bittner

    D. Meister et al. “A Survey on Bounding Volume Hierarchies for Ray Tracing”. en. In:Com- puter Graphics Forum40.2 (2021), pp. 683–712.doi:10 . 1111 / cgf . 142662.url:https : //onlinelibrary.wiley.com/doi/abs/10.1111/cgf.142662

    work page doi:10.1111/cgf.142662 2021

  12. [13]

    Geometrical Theory of Diffraction

    J. B. Keller. “Geometrical Theory of Diffraction”. In:Journal of the Optical Society of America52.2 (Feb. 1962), pp. 116–130.doi:10.1364/JOSA.52.000116

    work page doi:10.1364/josa.52.000116 1962

  13. [14]

    Asymptotic high-frequency methods

    R. Kouyoumjian. “Asymptotic high-frequency methods”. en. In:Proceedings of the IEEE53.8 (1965), pp. 864–876.doi:10.1109/PROC.1965.4065.url:http://ieeexplore.ieee.org/document/ 1445995/

    work page doi:10.1109/proc.1965.4065.url:http://ieeexplore.ieee.org/document/ 1965

  14. [15]

    Uniform Asymptotic Theory of Diffraction by a Plane Screen

    D. S. Ahluwalia, R. M. Lewis, and J. Boersma. “Uniform Asymptotic Theory of Diffraction by a Plane Screen”. en. In:SIAM Journal on Applied Mathematics16.4 (July 1968), pp. 783–807.doi: 10.1137/0116065.url:http://epubs.siam.org/doi/10.1137/0116065

    work page doi:10.1137/0116065.url:http://epubs.siam.org/doi/10.1137/0116065 1968

  15. [16]

    A uniform geometrical theory of diffraction for an edge in a perfectly conducting surface

    R. Kouyoumjian and P. Pathak. “A uniform geometrical theory of diffraction for an edge in a perfectly conducting surface”. en. In:Proceedings of the IEEE62.11 (1974), pp. 1448–1461.doi:10.1109/ PROC.1974.9651.url:https://ieeexplore.ieee.org/document/1451581

    work page arXiv 1974

  16. [17]

    P. Y. Ufimtsev.Fundamentals of the Physical Theory of Diffraction. Hoboken, NJ: John Wiley & Sons, 2007.doi:10.1002/0470109017

    work page doi:10.1002/0470109017 2007

  17. [18]

    Higher order equivalent edge currents for fringe wave radar scattering by perfectly con- ducting polygonal plates

    O. Breinbjerg. “Higher order equivalent edge currents for fringe wave radar scattering by perfectly con- ducting polygonal plates”. en. In:IEEE Transactions on Antennas and Propagation40.12 (Dec. 1992), pp. 1543–1554.doi:10.1109/8.204745.url:http://ieeexplore.ieee.org/document/ 204745/

    work page doi:10.1109/8.204745.url:http://ieeexplore.ieee.org/document/ 1992

  18. [19]

    In: Proc

    F. Tr ¨ager.Springer handbook of lasers and optics. en. New York: Springer, 2007.doi:10.1007/978- 0-387-30420-5. 13

    work page doi:10.1007/978- 2007

  19. [20]

    Peatross and M

    J. Peatross and M. Ware.Physics of light and optics. en. 2015 edition, February 3, 2023. Brigham Young University. Provo, Utah: Brigham Young University, Department of Physics, 2023

    work page 2015

  20. [21]

    C. A. Balanis.Balanis’ Advanced Engineering Electromagnetics. en. 1st ed. Wiley, Dec. 2023.doi: 10.1002/9781394180042.url:https://onlinelibrary.wiley.com/doi/book/10.1002/ 9781394180042

    work page doi:10.1002/9781394180042.url:https://onlinelibrary.wiley.com/doi/book/10.1002/ 2023

  21. [22]

    J. A. Stratton.Electromagnetic Theory. New York: McGraw-Hill, 1941.doi:10.1002/9781119134640

    work page doi:10.1002/9781119134640 1941

  22. [23]

    An Efficient Ray Tracing Algorithm for Multiple Straight Wedge Diffraction

    G. Carluccio and M. Albani. “An Efficient Ray Tracing Algorithm for Multiple Straight Wedge Diffraction”. en. In:IEEE Transactions on Antennas and Propagation56.11 (Nov. 2008), pp. 3534– 3542.doi:10.1109/TAP.2008.2005540.url:https://ieeexplore.ieee.org/document/ 4685916

    work page doi:10.1109/tap.2008.2005540.url:https://ieeexplore.ieee.org/document/ 2008

  23. [24]

    Resolution of Self-Intersection Issue in Monte Carlo Simulations Employing Graphics Ray Tracing Technology

    S. J. Kwon, J. Im, and H. G. Joo. “Resolution of Self-Intersection Issue in Monte Carlo Simulations Employing Graphics Ray Tracing Technology”. en. In: (2023). Korean Nuclear Society

    work page 2023

  24. [25]

    Pasquale.SagittaSBR.https://github.com/marco-pas/SagittaSBR

    M. Pasquale.SagittaSBR.https://github.com/marco-pas/SagittaSBR. 2026. 14

    work page 2026