rofunc.learning.RofuncML.utils

1.  Module Contents

1.1.  Functions

quaternion_from_homo_matrix	Return quaternion from rotation matrix.
get_canonical
multi_timestep_matrix
fd_transform	Finite difference transform matrix
multi_timestep_fd_q	param rs: list of std deviations of derivatives param xi_dim: param nb_past: param dt: return:
lifted_noise_matrix	Given a linear system with white noise, as in LQG,
lifted_transfer_matrix	Given a linear system
gu_pinv
create_relative_time	param q: [list of int] List of state indicator. ex: [0, 0, 0, 0, 1, 1, 1, 1, 2, 2, 2, 2, 2, 0, 0, 0, 1, 1, …] return time: [np.array(nb_timestep,)] Phase for each of the timestep
align_trajectories_hsmm
align_trajectories	param data: [list of np.array([nb_timestep, M, N, …])] return:
angle_to_rotation
feature_to_slice
dtype_to_index
plot_model_time
homo_matrix_from_quaternion	Return homogeneous rotation matrix from quaternion.
plot_demos_3d
repro_plot
plot_model
plot_demos
train_test
repro_demo

1.2.  API

rofunc.learning.RofuncML.utils.quaternion_from_homo_matrix(matrix, isprecise=False)

Return quaternion from rotation matrix.

If isprecise is True, the input matrix is assumed to be a precise rotation matrix and a faster algorithm is used.

rofunc.learning.RofuncML.utils.get_canonical(nb_dim, nb_deriv=2, dt=0.01)

rofunc.learning.RofuncML.utils.multi_timestep_matrix(A, B, nb_step=4)

rofunc.learning.RofuncML.utils.fd_transform(d, xi_dim, nb_past, dt=0.1)

Finite difference transform matrix

Parameters:

• d –

• xi_dim –

• nb_past –

• dt –

Returns:

rofunc.learning.RofuncML.utils.multi_timestep_fd_q(rs, xi_dim, dt)

Parameters:

• rs – list of std deviations of derivatives

• xi_dim –

• nb_past –

• dt –

Returns:

rofunc.learning.RofuncML.utils.lifted_noise_matrix(A=None, B=None, nb_dim=3, dt=0.01, horizon=50)

Given a linear system with white noise, as in LQG,

\[\xi_{t+1} = \mathbf{A} (\xi_t + w_i) + \mathbf{B} u_t + v_i\]

returns the lifted form for noise addition, s_v, s_w,

\[\mathbf{\xi} = \mathbf{S}_{\xi} \xi_0 + \mathbf{S}_u \mathbf{u} + \mathbf{S}_v + \mathbf{S}_w\]

Returns:

s_u

rofunc.learning.RofuncML.utils.lifted_transfer_matrix(A=None, B=None, nb_dim=3, dt=0.01, horizon=50, sparse=False)

Given a linear system

\[\xi_{t+1} = \mathbf{A} \xi_t + \mathbf{B} u_t\]

returns the lifted form for T timesteps

\[\mathbf{\xi} = \mathbf{S}_{\xi} \xi_0 + \mathbf{S}_u \mathbf{u}\]

rofunc.learning.RofuncML.utils.gu_pinv(A, rcond=1e-15)

rofunc.learning.RofuncML.utils.create_relative_time(q, start=-1.0)

Parameters:

q – [list of int] List of state indicator. ex: [0, 0, 0, 0, 1, 1, 1, 1, 2, 2, 2, 2, 2, 0, 0, 0, 1, 1, …]

Return time:

[np.array(nb_timestep,)] Phase for each of the timestep

rofunc.learning.RofuncML.utils.align_trajectories_hsmm(data, nb_states=5)

rofunc.learning.RofuncML.utils.align_trajectories(data, additional_data=[], hsmm=True, nb_states=5)

Parameters:

data – [list of np.array([nb_timestep, M, N, …])]

Returns:

rofunc.learning.RofuncML.utils.angle_to_rotation(theta)

rofunc.learning.RofuncML.utils.feature_to_slice(nb_dim=2, nb_frames=None, nb_attractor=2, features=None)

rofunc.learning.RofuncML.utils.dtype_to_index(dtype)

rofunc.learning.RofuncML.utils.plot_model_time(model, demos, figsize=(10, 2), dim_idx=[1], demo_idx=0)

rofunc.learning.RofuncML.utils.homo_matrix_from_quaternion(quaternion)

Return homogeneous rotation matrix from quaternion.

>>> M = homo_matrix_from_quaternion([0.99810947, 0.06146124, 0, 0])
>>> np.allclose(M, rotation_matrix(0.123, [1, 0, 0]))
True
>>> M = homo_matrix_from_quaternion([1, 0, 0, 0])
>>> np.allclose(M, np.identity(4))
True
>>> M = homo_matrix_from_quaternion([0, 1, 0, 0])
>>> np.allclose(M, np.diag([1, -1, -1, 1]))
True

rofunc.learning.RofuncML.utils.plot_demos_3d(demos, figsize=(15, 5), angle=[60, 45])

rofunc.learning.RofuncML.utils.repro_plot(model, demos, save=False, tp_list=[], figsize=(3.5, 5))

rofunc.learning.RofuncML.utils.plot_model(model, demos, figsize=(8, 3.5), skill_name='temp', save=False)

rofunc.learning.RofuncML.utils.plot_demos(demos, data_dim, figsize=(8, 5))

rofunc.learning.RofuncML.utils.train_test(demos, demo_idx=0, nb_states=5, test=True, sensory=True, kbins=True, hmmr=True, nb_dim=3, nb_frames=2)

rofunc.learning.RofuncML.utils.repro_demo(model, hmmr, demos, demo_idx=0, start_point=None, plot_on=False)