InverseKinematics

class bayesflow.simulators.InverseKinematics(scales: ndarray = None, l1: float = 0.5, l2: float = 0.5, l3: float = 1.0, rng: Generator = None)

Bases: BenchmarkSimulator

Inverse Kinematics simulated benchmark See: https://arxiv.org/pdf/2101.10763.pdf

Parameters:

scales: np.ndarray of shape (4, ) or None, optional, default: None

The four scales of the Gaussian prior. If None provided, the scales from https://arxiv.org/pdf/2101.10763.pdf will be used: [0.25, 0.5, 0.5, 0.5].

l1: float, optional, default: 0.5

The length of the first segment.

l2: float, optional, default: 0.5

The length of the second segment.

l3: float, optional, default: 1.0

The length of the third segment.

rng: np.random.Generator or None, optional, default: None

An optional random number generator to use.

__call__(**kwargs) → dict[str, ndarray]

Call self as a function.

prior()

Generates a random draw from a 4-dimensional Gaussian prior distribution with a spherical convariance matrix. The parameters represent a robot’s arm configuration, with the first parameter indicating the arm’s height and the remaining three are angles.

Returns:

paramsnp.ndarray of shape (4, )

A single draw from the 4-dimensional Gaussian prior.

rejection_sample(batch_shape: tuple[int, ...], predicate: Callable[[dict[str, ndarray]], ndarray], *, axis: int = 0, sample_size: int = None, **kwargs) → dict[str, ndarray]

sample(batch_shape: tuple[int, ...], **kwargs) → dict[str, ndarray]

Runs simulated benchmark and returns batch_size parameter and observation batches

Parameters:

batch_shape: tuple

Number of parameter-observation batches to simulate.

Returns:

dict[str, np.ndarray]: simulated parameters and observables

with shapes (batch_size, …)

sample_batched(batch_shape: tuple[int, ...], *, sample_size: int, **kwargs)

Sample the desired number of simulations in smaller batches.

Limited resources, especially memory, can make it necessary to run simulations in smaller batches. The number of samples per simulated batch is specified by sample_size.

Parameters:

batch_shapeShape

The desired output shape, as in sample(). Will be rounded up to the next complete batch.

sample_sizeint

The number of samples in each simulated batch.

kwargs

Additional keyword arguments passed to sample().

observation_model(params: ndarray)

Returns the 2D coordinates of a robot arm given parameter vector. The first parameter represents the arm’s height and the remaining three correspond to angles.

Parameters:

paramsnp.ndarray of shape (params, )

The four model parameters which will determine the coordinates

Returns:

xnp.ndarray of shape (2, )

The 2D coordinates of the arm