/DRL_robotics

I apply DDPG+HER for a Nachi_mz07 robot arm with ROS, Gazebo and OpenAI_ROS framework

Primary LanguagePythonMIT LicenseMIT

DDPG and HER for robotics

I apply DDPG+HER for a Nachi_mz07 robot arm with ROS, Gazebo and OpenAI_ROS framework My work is the extension of the work of Tianhong Dai: https://github.com/TianhongDai/hindsight-experience-replay

New feature:

I changed the output from torque value to delta position value. I want the robot to learn how to move based on position controller instead of torque controller. It's more safety in an industrial application.

I changed the HER buffer in TianhongDai/OpenAI work from a static to dynamics size, it is now applicable to the case of early ending episode

My System

  • ROS Noetic
  • Python 3.8
  • Pytorch 1.12.0 + Cu116
  • Ubuntu 20.04.5 LTS focal
  • Gym 0.21.0
  • MoveIt 1

Demo video:

Instruction:

  • Will be added later