Training environments

A number of custom Gym environments are available in the gym_envs directory. They simulate the WidowX MK-II robotic manipulator with the Pybullet physics engine. The objective is to bring the end-effector as close as possible to a target position.

Each implemented environment is described here. The action, observation and reward functions are given in this document. Various types of reaching tasks are available. They are listed in below by increasing order of difficulty.

Reaching tasks available

Reaching task

Rendering

_images/fixed_pos_no_orient.gif

Fixed position + no orientation

_images/rand_pos_no_orient.gif

Random position + no orientation

_images/fixed_pos_fixed_orient.gif

Fixed position + fixed orientation

_images/fixed_pos_rand_orient.gif

Fixed position + random orientation

_images/moving_pos_no_orient.gif

Moving position + no orientation