errorsimulator.py

class EErrType(value)

Bases: Enum

Enumeration defining the error type for separation of error types for later analysis.

EErrType.SYSTEMATIC:

Also known as an epistemic error and is due to a lack of knowledge. Common examples include spatial or temporal averaging, digitisation / round off error and calibration errors.

EErrType.RANDOM:

Also known as aleatory error and is generally a result of sensor noise.

SYSTEMATIC = 1

RANDOM = 2

class EErrDep(value)

Bases: Enum

Enumeration defining error dependence.

EErrDep.INDEPENDENT:

Errors are calculated based on the ground truth sensor values interpolated from the input simulation.

EErrDep.DEPENDENT:

Errors are calculated based on the accumulated sensor reading due to all preceeding errors in the chain.

INDEPENDENT = 1

DEPENDENT = 2

class IErrSimulator

Bases: ABC

Interface (abstract base class) for sensor error simulation allowing a list of simulated errors (i.e. an error chain) to be constructed and executed in order.

abstractmethod get_error_type()

Gets the error type enumeration as either random or systematic. Random errors sample at every time step and systematic errors typically apply a bias that is constant over time.

Returns:

EErrType – Enumeration definining RANDOM or SYSTEMATIC error types.

abstractmethod get_error_dep()

Gets the error dependence enumeration value. Independent errors are calculated based on the ground truth and ignore other errors in the error chain. Dependent errors are calculated based on the accumulated measurement value at their place in the error chain.

Returns:

EErrDep – Enumeration definining INDEPENDENT or DEPENDENT error dependence.

abstractmethod set_error_dep(dependence)

Sets the error dependence for errors that support changing the dependence. Independent errors are calculated based on the ground truth and ignore other errors in the error chain. Dependent errors are calculated based on the accumulated measurement value at their place in the error chain.

Parameters:

dependence (EErrDep) – Enumeration definining INDEPENDENT or DEPENDENT error dependence.

abstractmethod reseed(seed=None)

Reseeds the random generators of the error simulator. Mainly used for multi-processed simulations which inherit the same seed as the main process so need to be reseeded. If the error simulator does not have any random generators then this function implementation will be empty.

Parameters:

seed (int | None, optional) – Integer seed for the random number generator, by default None. If None then the seed is generated using OS entropy (see numpy docs).

abstractmethod sim_errs(err_basis, sens_data)

Creates the simulated error array based on the input error basis array. The output error array will be the same shape as the input error basis array.

Parameters:

• err_basis (np.ndarray) – Used as the base array for calculating the returned error. If the error is independent this will be the ‘truth’ array and if the error is dependent this will be the accumulated sensor measurement array at this point in the error chain.

• sens_data (SensorData) – Sensor data object holding the current sensor state before applying this error calculation.

Returns:

tuple[np.ndarray, SensorData] – Tuple containing the error array from this calculator and a SensorData object with the current accumulated sensor state starting from the nominal state up to and including this error calculator in the error chain. Note that many errors do not modify the sensor data so the sensor data class is passed through this function unchanged.