chenlin/Mujoco_WASM
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Replace SdfLib with linear octree interpolation of TriangleMeshDistance.

Browse Source 
Temporary changes to the octree:
- Changed frame from mesh to geom

Before change (tolerance 1e-3):
```
 Simulation time      : 2.34 s
 Steps per second     : 4275
 Realtime factor      : 8.55 x
 Time per step        : 233.9 µs

 Newton iters / step  : 2.26
 Contacts / step      : 3.69
 Constraints / step   : 14.76
 Degrees of freedom   : 12
 Dynamic memory usage : 0.7% of 14M
```

After change (6 octree levels):
```
 Simulation time      : 2.24 s
 Steps per second     : 4472
 Realtime factor      : 8.94 x
 Time per step        : 223.6 µs

 Newton iters / step  : 2.47
 Contacts / step      : 3.37
 Constraints / step   : 13.49
 Degrees of freedom   : 12
 Dynamic memory usage : 0.2% of 14M
```

PiperOrigin-RevId: 781019754
Change-Id: Ib15581244dfe9e571e1c6c2cac4a101cf3d8ba3d
This commit is contained in:
Alessio Quaglino
2025-07-09 06:20:36 -07:00
committed by Copybara-Service
parent e6c5715903
commit b81f1db8af
8 changed files with 194 additions and 121 deletions
Download Patch File Download Diff File Expand all files Collapse all files
plugin/sdf/README.md
+5 -5
View File
@@ -66,11 +66,11 @@ Parameters:
Implemented in [sdflib.cc](sdflib.cc). Example usage in [cow.xml](../../model/plugin/sdf/cow.xml).
This plugin uses the library [SdfLib](https://github.com/UPC-ViRVIG/SdfLib) to compute a voxel-based approximation of a
user-specified mesh. The mesh can be arbitrary and not necessarily convex. This offers an alternative to
convex-decomposed meshes. The performance is likely to be slower than that of analytic SDFs, since a cubic
approximation has to be evaluated on the convex grid. However, the SDF generation is done automatically, simplifying the
task of creating an SDF, which can be difficult for complex shapes.
This plugin uses the library [TriangleMeshDistance](https://github.com/InteractiveComputerGraphics/TriangleMeshDistance)
to compute a voxel-based approximation of a user-specified mesh. The mesh can be arbitrary and not necessarily convex.
This offers an alternative to convex-decomposed meshes. The performance is likely to be slower than that of analytic
SDFs, since a cubic approximation has to be evaluated on the convex grid. However, the SDF generation is done
automatically, simplifying the task of creating an SDF, which can be difficult for complex shapes.
### How to make your own SDF