Hi Andy,
I tried looking for inbuilt methods for FDTD simulations, but it seems like it is there are none as of now. Like you mentioned, you can extract the points in 3D from scenario.Meshes. I’ve created a script which concatenates points from all buildings:
% Create the UAV scene as you have provided
% Initialize containers for vertices and faces across all meshes
allVertices = [];
allFaces = [];
% Initialize counters for vertices and faces
vertexCount = 0;
% Loop through the building meshes in the scene
for meshIdx = 1:length(scene.Meshes)
% for meshIdx = 1:1
% Extract vertices and faces from the current mesh
vertices = scene.Meshes{meshIdx}.Vertices;
faces = scene.Meshes{meshIdx}.Faces;
% Check if adjustment is needed for the face indices
if vertexCount > 0
% Adjust face indices based on the current vertex count
faces = faces + vertexCount;
end
% Append to the allVertices and allFaces containers
allVertices = [allVertices; vertices];
allFaces = [allFaces; faces];
% Update the vertex count
vertexCount = size(allVertices, 1);
end
The above script also combines the face (triangle connectivity matrix) data with which you can use view the 3D meshes. You can use trimesh function to view the grid:
trimesh(allFaces, allVertices(:,1),allVertices(:,2),allVertices(:,3))
Which will give the following output:
If the points are your only requirement, you can use the allVertices variable.
But I’ve also tried an alternative method to FDTD by modelling the above mesh using createpde, which can be used for electromagnetic analyses, but it requires the mesh to have closed surfaces. Hence, it will require removal of the ground plane (the first mesh) by changing the for loop to:
for meshIdx = 2:length(scene.Meshes)
Then we get the following grid:
Now the vertices and faces can be imported as a mesh, and a model can be created and solved as follows:
% Create a PDE model for electromagnetic wave analysis
model = createpde('electromagnetic', 'electrostatic');
% Create geometry from mesh and append to the PDE model
geometryFromMesh(model, allVertices', allFaces');
% Display the geometry
figure(2);
pdegplot(model);
title('Geometry Imported into PDE Model');
% Define electromagnetic properties for the buildings
electromagneticProperties(model, 'RelativePermittivity', 4, 'RelativePermeability', 1, 'Conductivity', 0.01);
% Apply boundary condition
bc = electromagneticBC(model, 'Face', 1:model.Geometry.NumFaces, 'Voltage', 0);
model.VacuumPermittivity = 8.8541878128e-12;
% Define ChargeDensity source
sc1 = electromagneticSource(model,"Cell",1,"ChargeDensity",0.3)
% Generate mesh
generateMesh(model); % Adjust 'Hmax' based on the desired resolution
% Solve the PDE
result = solve(model);
% Plot the solution
figure(3);
pdeplot3D(model, 'ColorMapData', abs(result.NodalSolution));
title('Field Distribution');
The above is just a sample electromagnetic scenario which you can modify. To get results across time-domain, you can modify and solve the model for each timestep. This could get computationally expensive, but the solution can be parallelized using parfor (https://www.mathworks.com/help/parallel-computing/parfor.html), as solutions for each timestep are independent of each other.
Here are documentation links for some of the functions I used:
- trimesh (to plot a triangular mesh) - https://www.mathworks.com/help/matlab/ref/trimesh.html
- createpde (for creating an electromagnetic model) - https://www.mathworks.com/help/pde/ug/createpde.html
- electromagneticBC (define boundary conditions) - https://www.mathworks.com/help/pde/ug/pde.electromagneticmodel.electromagneticbc.html
- electromagneticSource (specify current density, charge density, or magnetization values) - https://www.mathworks.com/help/pde/ug/pde.electromagneticmodel.electromagneticsource.html
I hope this helps.
