def cmu_humanoid_run_walls(random_state=None):
"""Requires a CMU humanoid to run down a corridor obstructed by walls."""
# Build a position-controlled CMU humanoid walker.
walker = cmu_humanoid.CMUHumanoidPositionControlled(
observable_options={'egocentric_camera': dict(enabled=True)})
# Build a corridor-shaped arena that is obstructed by walls.
arena = corr_arenas.WallsCorridor(wall_gap=4.,
wall_width=distributions.Uniform(1, 7),
wall_height=3.0,
corridor_width=10,
corridor_length=100,
include_initial_padding=False)
# Build a task that rewards the agent for running down the corridor at a
# specific velocity.
task = corr_tasks.RunThroughCorridor(walker=walker,
arena=arena,
walker_spawn_position=(0.5, 0, 0),
target_velocity=3.0,
physics_timestep=0.005,
control_timestep=0.03)
return composer.Environment(time_limit=30,
task=task,
random_state=random_state,
strip_singleton_obs_buffer_dim=True)
def ant_run_walls():
walker = ant.Ant()
arena = corr_arenas.WallsCorridor(wall_gap=4.,
wall_width=distributions.Uniform(1, 7),
wall_height=3.0,
corridor_width=10,
corridor_length=250,
include_initial_padding=False)
task = corr_tasks.RunThroughCorridor(walker=walker,
arena=arena,
walker_spawn_position=(0.5, 0, 0),
walker_spawn_rotation=0,
stand_height=0.2,
contact_termination=False,
physics_timestep=_PHYSICS_TIMESTEP,
control_timestep=_CONTROL_TIMESTEP)
# (Chongyi Zheng): redefine reward function
# task.get_reward = _ant_get_reward.__get__(task, task.get_reward)
return composer.Environment(time_limit=30,
task=task,
strip_singleton_obs_buffer_dim=True)
def jumping_ball_run_walls():
walker = jumping_ball.JumpingBallWithHead()
arena = corr_arenas.WallsCorridor(wall_gap=4.,
wall_width=distributions.Uniform(1, 7),
wall_height=3.0,
corridor_width=10,
corridor_length=250,
include_initial_padding=False)
task = corr_tasks.RunThroughCorridor(walker=walker,
arena=arena,
walker_spawn_position=(0.5, 0, 0),
walker_spawn_rotation=0,
contact_termination=False,
physics_timestep=_PHYSICS_TIMESTEP,
control_timestep=_CONTROL_TIMESTEP)
return composer.Environment(time_limit=30,
task=task,
strip_singleton_obs_buffer_dim=True)
def ant_run_walls(random_state=None):
walker = ant.Ant()
arena = corr_arenas.WallsCorridor(wall_gap=4.,
wall_width=distributions.Uniform(1, 7),
wall_height=3.0,
corridor_width=10,
corridor_length=100,
include_initial_padding=False)
task = corr_tasks.RunThroughCorridor(walker=walker,
arena=arena,
walker_spawn_position=(0.5, 0, 0),
walker_spawn_rotation=0,
physics_timestep=_PHYSICS_TIMESTEP,
control_timestep=_CONTROL_TIMESTEP)
return composer.Environment(time_limit=30,
task=task,
random_state=random_state,
strip_singleton_obs_buffer_dim=True)