def test_real_env_contructors(dt, max_steps):
qbbr = QBallBalancerReal(dt=dt, max_steps=max_steps)
assert qbbr is not None
qcp_st = QCartPoleStabReal(dt=dt, max_steps=max_steps)
assert qcp_st is not None
qcp_su = QCartPoleSwingUpReal(dt=dt, max_steps=max_steps)
assert qcp_su is not None
qqr = QQubeReal(dt=dt, max_steps=max_steps)
assert qqr is not None
if __name__ == '__main__':
# Parse command line arguments
args = get_argparser().parse_args()
# Get the experiment's directory to load from
ex_dir = ask_for_experiment() if args.ex_dir is None else args.ex_dir
ex_tag = ex_dir.split('--', 1)[1]
# Load the policy and the environment (for constructing the real-world counterpart)
env_sim, policy, _ = load_experiment(ex_dir, args)
if args.verbose:
print(f'Policy params:\n{policy.param_values.detach().cpu().numpy()}')
# Create real-world counterpart (without domain randomization)
env_real = QBallBalancerReal(args.dt)
print_cbt('Set up the QBallBalancerReal environment.', 'c')
# Set up PD-controller
pdctrl = QBallBalancerPDCtrl(env_real.spec)
if args.random_init_state:
# Random initial state
min_init_state = np.array([0.7 * 0.275 / 2, 0])
max_init_state = np.array([0.8 * 0.275 / 2, 2 * np.pi])
init_space = Polar2DPosSpace(min_init_state, max_init_state)
init_state = init_space.sample_uniform()
else:
# Center of the plate is initial state
# init_state = np.zeros(2)
# init_state = np.array([1/np.sqrt(2), 1/np.sqrt(2)])
# init_state = np.array([1., 0.])
from pyrado.utils.input_output import print_cbt
from pyrado.utils.argparser import get_argparser
if __name__ == '__main__':
# Parse command line arguments
args = get_argparser().parse_args()
# Get the experiment's directory to load from
ex_dir = ask_for_experiment()
# Load the policy (trained in simulation) and the environment (for constructing the real-world counterpart)
env_sim, policy, _ = load_experiment(ex_dir)
# Detect the correct real-world counterpart and create it
if isinstance(inner_env(env_sim), QBallBalancerSim):
env_real = QBallBalancerReal(dt=args.dt, max_steps=args.max_steps)
elif isinstance(inner_env(env_sim), QCartPoleSim):
env_real = QCartPoleReal(dt=args.dt, max_steps=args.max_steps)
elif isinstance(inner_env(env_sim), QQubeSim):
env_real = QQubeReal(dt=args.dt, max_steps=args.max_steps)
else:
raise pyrado.TypeErr(
given=env_sim,
expected_type=[QBallBalancerSim, QCartPoleSim, QQubeSim])
print_cbt(f'Set up env {env_real.name}.', 'c')
# Finally wrap the env in the same as done during training
env_real = wrap_like_other_env(env_real, env_sim)
# Run on device
done = False
default=2e-5,
help='Step size for the action')
parser.add_argument(
'--thold_theta_deg',
type=float,
default=0.1,
help='Threshold for the servo shafts to rotate in degree')
parser.add_argument('--thold_x',
type=float,
default=5e-4,
help='Threshold for the cart to move in meter')
args = parser.parse_args()
if args.env_name == QBallBalancerReal.name:
# Create the interface for communicating with the device
env = QBallBalancerReal(args.dt, args.max_steps)
print_cbt('Set up the QBallBalancerReal environment.', 'c')
# Initialize
thold_theta = args.thold_theta_deg / 180. * np.pi
tholds_V = np.zeros(
(args.num_runs, 2,
2)) # num_runs matrices in format [[-x, +x], [-y, +y]]
env.reset() # sends a zero action until sth else is commanded
input(
'Servo voltages set to [0, 0]. Prepare and then press enter to start.'
)
# Run
for r in range(args.num_runs):
Returns:
The bounded variable.
"""
return np.maximum(min_value, np.minimum(x, max_value))
# python ... -env-name qcp-st
if __name__ == '__main__':
# Parse command line arguments
args = get_argparser().parse_args()
# Set up PD-controller
if args.env_name in QBallBalancerReal.name:
env = QBallBalancerReal(args.dt, args.max_steps)
policy = QBallBalancerPDCtrl(env.spec,
kp=to.diag(to.tensor([3.45, 3.45])),
kd=to.diag(to.tensor([2.11, 2.11])))
print_cbt('Set up controller for the QBallBalancerReal environment.',
'c')
elif args.env_name == QCartPoleStabReal.name:
env = QCartPoleStabReal(args.dt, args.max_steps)
policy = CartpoleStabilizerPolicy(env.spec,
K=np.array([
1.2278416e+00, 4.5279346e+00,
-1.2385756e-02, 6.0038762e+00,
-4.1818547e+00
]))
# policy = QCartPoleSwingUpAndBalanceCtrl(env.spec)
def default_qqbb_real():
return QBallBalancerReal(dt=1 / 500.0, max_steps=500)