Chris Paxton's Projects
Action Chunking Transformer implementation for low cost robot
Barrett WAM & BHand-280 CAD and modular URDFs with inertial properties.
A lightweight implementation of Behavior Trees in Python that can interface with the ROS framework.
Teaching robots to respond to open-vocab queries with CLIP and NeRF-like neural fields
Efficient 6-DoF Grasp Generation in Cluttered Scenes
Example workspace for the CoSTAR project, containing YAML files describing data for tasks that can be performed via ROS.
Helpful gazebo plugin that exports a moveit planning scene for an arbitrary gazebo simulation.
User interface for behavior tree based authoring of robot task plans.
Integrated ROS capabilities for planning, predicate inference, gripper control, and perception for use with the KUKA LBR IIWA and Universal Robots.
Implementation of CoW appearing at CVPR 2023
Website
moveit packages code maintain for Crops
This ROS package is a general, robot-agnostic implementation of dynamic movement primitives (DMPs).
Action server for DMP replay
Gesture learning and execution via DMPs
Dobb·E: An open-source, general framework for learning household robotic manipulation
daVinci Research Kit ROS stack
A modular embodied agent architecture and platform for building embodied agents
Stack trace visualizer
Wrappers, tools and additional API's for using ROS with Gazebo. Formally simulator_gazebo stack
Geometric Control Optimization and Planning Library
Graphical entity-relationship model
Baseline model for "GraspNet-1Billion: A Large-Scale Benchmark for General Object Grasping" (CVPR 2020)
MATLAB code for learning primitive actions from human demonstrations and using these for planning complex tasks, applied to levels of the Needle Master android game.
Integrated task and motion planning using grounded actions learned from observing expert demonstrations.
Habitat2 compatible version of the Hello Robot Stretch urdf
A modular high-level library to train embodied AI agents across a variety of tasks, environments, and simulators.