1. [Publications](/publications)
2. Object Rearrangement Using Learned Implicit Collision Functions
 
 # Object Rearrangement Using Learned Implicit Collision Functions

  ![](/sites/default/files/styles/wide/public/publications/rearrangement_v3.png?itok=b-fnVcUd)

 Robotic object rearrangement combines the skills of picking and placing objects. When object models are unavailable, typical collision-checking models may be unable to predict collisions in partial point clouds with occlusions, making generation of collision-free grasping or placement trajectories challenging. We propose a learned collision model that accepts scene and query object point clouds and predicts collisions for 6DOF object poses within the scene. We train the model on a synthetic set of 1 million scene/object point cloud pairs and 2~billion collision queries. We leverage the learned collision model as part of a model predictive path integral (MPPI) policy in a tabletop rearrangement task and show that the policy can plan collision-free grasps and placements for objects unseen in training in both simulated and physical cluttered scenes with a Franka Panda robot. The learned model outperforms both traditional pipelines and learned ablations by 9.8% in accuracy on a dataset of simulated collision queries and is 75x faster than the best-performing baseline.



 ## Authors



Michael Danielczuk (UC Berkeley)

Arsalan Mousavian (NVIDIA)

Clemens Eppner (NVIDIA)

Dieter Fox (NVIDIA)

 

 

 ## Publication Date



Wednesday, March 24, 2021

 

 ## Published in



[ICRA 2021](http://www.icra2021.org/)

 

 ## Research Area



[Robotics](/research-area/robotics)

 

 

 ## External Links



[Paper](https://arxiv.org/pdf/2011.10726.pdf)

[Video Presentation](https://youtu.be/SHC2ODD1QcU)

[Video](https://youtu.be/anXPw7o7Wx8)

[Code \[Coming Soon\]](https://github.com/NVlabs/scene_collisionnet)

[Experiment Videos](https://sites.google.com/nvidia.com/scenecollisionnet)