Differentiable Robotics

Collection of differentiable methods for robotics applications implemented with Pytorch.

Robot-terrain interaction

cd scripts/
python robot_terrain_interaction.py

Reference: MonoForce: Self-supervised Learning of Physics-informed Model for Predicting Robot-terrain Interaction

ICP point cloud alignment

cd src/
python -m icp.align_clouds

Reference: Self-Supervised Depth Correction of Lidar Measurements From Map Consistency Loss

Fit a bounding box to a point cloud

The optimization is based on a point cloud coverage function by the bounding box.

Path smoothing

cd src/
python -m trajopt.smooth_path