def start(self, robot_name: str, fsm_config: str = 'SingleRun') -> None:
self.output_dir = f'test_dir/{get_filename_without_extension(__file__)}'
os.makedirs(self.output_dir, exist_ok=True)
config_dict = {
'output_path': self.output_dir,
'factory_key': "ROS",
'max_number_of_steps': -1,
'ros_config': {
'ros_launch_config': {
'control_mapping_config': 'keyboard_and_joystick',
'fsm_mode': fsm_config,
'gazebo': 'sim' in robot_name,
'random_seed': 123,
'robot_name': robot_name,
'world_name': 'empty',
'robot_display': True,
},
'actor_configs': [{
'file': f'src/sim/ros/config/actor/keyboard_drone_sim.yml',
'name': 'keyboard'
}, {
'file': f'src/sim/ros/config/actor/joystick_drone_sim.yml',
'name': 'joystick'
}],
'visible_xterm':
True,
},
}
config = EnvironmentConfig().create(config_dict=config_dict)
self._environment = RosEnvironment(config=config)
def setUp(self) -> None:
config_dict['output_path'] = f'test_dir/{get_filename_without_extension(__file__)}'
config = EnvironmentConfig().create(
config_dict=config_dict
)
self.output_dir = config.output_path
self._environment = RosEnvironment(
config=config
)
def setUp(self) -> None:
self.output_dir = f'test_dir/{get_filename_without_extension(__file__)}'
os.makedirs(self.output_dir, exist_ok=True)
config_dict['output_path'] = self.output_dir
config = EnvironmentConfig().create(
config_dict=config_dict
)
self._environment = RosEnvironment(
config=config
)
def main():
global environment
config_file = Parser().parse_args().config
config = EnvironmentConfig().create(config_file=config_file)
signal.signal(signal.SIGTERM, signal_handler)
# try:
environment = RosEnvironment(config=config)
state = environment.reset()
while state.done != TerminationType.Success and state.done != TerminationType.Failure:
print(f'state: {environment.fsm_state}')
state = environment.step()
environment.remove()
def _test_environment_by_name(self, name: str) -> bool:
config = {
'output_path': self.output_dir,
'factory_key': "GYM",
'max_number_of_steps': 200,
'ros_config': None,
'gym_config': {
'random_seed': 123,
'world_name': name,
'render': False
}
}
environment = GymEnvironment(
EnvironmentConfig().create(config_dict=config))
experience, _ = environment.reset()
while experience.done != TerminationType.Done:
experience, _ = environment.step(environment.get_random_action())
return environment.remove() == ProcessState.Terminated
"python" if DS_TASK == "velocities" else "mathias_controller",
"world_name": WORLD,
"starting_height": 1.5,
"z_pos": 0.2,
"yaw_or": 0.0,
"gazebo": True,
},
"actor_configs": [{
"name":
"mathias_controller_with_KF",
"file":
f'{os.environ["CODEDIR"]}/src/sim/ros/config/actor/mathias_controller_with_KF.yml',
}],
},
}
config = EnvironmentConfig().create(config_dict=environment_config_dict)
output_dir = config.output_path
# Load model
model = fgbg.DownstreamNet(
output_size=(4, ) if DS_TASK == "velocities" else (3, ),
batch_norm=BATCHNORM)
ckpt = torch.load(CHECKPOINT + "/checkpoint_model.ckpt",
map_location=torch.device("cpu"))
model.load_state_dict(ckpt["state_dict"])
print(f"Loaded encoder from {CHECKPOINT}.")
environment = RosEnvironment(config=config)
losses = []
successes = []