Skip to content

AEGIS

Adaptive Electromagnetic Geometric Illumination & Safety

AEGIS computes absorbed power density on human body surfaces in wireless environments. It replaces volumetric EM simulation (\(10^{12}\) voxels) with \(O(MN)\) surface operations by exploiting the geometric nature of mmWave dosimetry.

The core equation:

\[S_{\mathrm{ab}}(\mathbf{r}) = S_{\mathrm{inc}} \cdot T_0 \cdot [\hat{n}(\mathbf{r}) \cdot (-\hat{k})]_+\]

Nine fidelity levels (0-8) provide a controlled accuracy-cost tradeoff, from \(O(1)\) worst-case bounds to \(O(M_{\mathrm{ant}}^3)\) exposure-constrained beamforming.

Absorbed power density on a human body phantom

  • Getting started

    Install AEGIS and run your first dosimetry computation in under a minute.

    Installation

  • Fidelity levels

    Nine levels from worst-case bounds to coherent MIMO beamforming.

    Levels 0-8

  • Tissue and geometry

    Cole-Cole dielectric models, Fresnel transmission, and body mesh operations.

    User guide

  • API reference

    Auto-generated from source. Every public class and function documented.

    Reference

  • Architecture

    Module structure, data flow, and design principles.

    Developer guide

  • Testing

    Golden tests, property tests, Mie regression, and 260+ test cases.

    Test guide

Quick start

from aegis import DosimetryEngine, BodyMesh, TissueModel, PropagationPaths

skin = TissueModel.from_database("Skin", 28e9)
body = BodyMesh.load("thelonious.stl")
paths = PropagationPaths.from_powers(k_hat=[[0, 0, -1]], power=[1.0])

engine = DosimetryEngine(skin)
result = engine.compute(body, paths, level=2)

print(f"P_abs = {result.p_abs:.4f} W")
print(f"Peak S_ab = {result.peak_sab:.2f} W/m²")

How it works

AEGIS treats the human body as a triangle mesh and incoming wireless signals as propagation paths (directions + powers). For each triangle, the kernel computes absorbed power density based on the angle between the surface normal and the incoming wave direction.

The key insight: at mmWave frequencies, the skin depth is so shallow (< 0.5 mm) that absorption is entirely a surface phenomenon. This makes the \(O(MN)\) geometric computation exact to within 0.35% of the full Fresnel solution.

Fidelity levels at a glance

Level Name What it adds Cost
0 Bound Worst-case \(P_{\mathrm{abs}}\) \(O(1)\)
1 Aggregate SH-compressed directivity \(O(L^2)\)
2 Geometric ReLU kernel on mesh \(O(MN)\)
3 Fresnel Angle-dependent \(T(\theta)\) \(O(MN)\)
4 Polarisation TE/TM decomposition \(O(MN)\)
5 Curvature Local curvature correction \(O(MN)\)
6 Diffraction GELU shadow smoothing \(O(MN)\)
7 Coherent Complex field summation \(O(MNK)\)
8 ECBF Exposure-constrained beamforming \(O(K^3)\)

In practice, all nine levels complete in milliseconds for typical meshes. Pick the level that matches your physics requirements, not your performance budget.

WAVES Ghent University imec