Gym-Styled Panda and UR5 with Robotiq-85 / 140 gripper in PyBullet simulator

This repo is under active development.

Highlights

UR5 arm with end-effector 6D IK (Position [X Y Z] and Orientation [R P Y])
Enhanced Robotiq-85 / 140 gripper, with precise position control and experimental torque control
Built-in YCB models loader (and obj models after decomposition)
Gym-styled API, making it suitable for Reinforcement Learning in the field of push-and-grasp
A heuristic grasping demo
An interactive user-control demo

You can try this repo with the interactive user-control demo.

python main.py

You can change the XYZ and RPY of the gripper on the right panel.

Press Z to close the gripper and R to open it.

See heuristic demo in main.py.