env.md

Environment

If you want to use our environment for your research, you can check this README file.

Try our environment with random action

We provide several examples in the example directory to try our environments, including relocate, pour, and place inside. It is very simple to use the code, take pour as example:

import numpy as np

from hand_imitation.env.environments.mug_pour_water_env import WaterPouringEnv

if __name__ == '__main__':
    # Environment parameters
    tank_size = (0.15, 0.15, 0.06)  # the size of water tank container
    mug_init_offset = (0.22, 0)  # the initial position of mug
    tank_init_pos = (-0.08, -0.1)  # the initial position of the water tank container

    # Geom parameters for the MuJoCo mug object
    # Please refer to http://www.mujoco.org/book/XMLreference.html#geom for more details
    geom_params = dict(condim="4", margin="0.003")

    # Construct env
    env = WaterPouringEnv(has_renderer=True, tank_size=tank_size, mug_init_offset=mug_init_offset,
                          tank_init_pos=tank_init_pos, **geom_params)
    # or you can simply use the default parameters with:
    # env = WaterPouringEnv(has_renderer=True)
    env.seed(0)

    # Action spec
    obs = env.reset()
    low, high = env.action_spec

    # Visualization
    for i in range(500):
        action = np.random.uniform(low, high)
        obs, reward, done, _ = env.step(action)
        env.render()