BinaryMeasurement - GTSAM Docs

Created by Codex.

Store a typed measurement between two keys without defining an error function.

Authors: Frank Dellaert, et al. (see THANKS for the full author list)

See LICENSE for the license information

try:
    import google.colab
    %pip install --quiet gtsam-develop
except ImportError:
    pass

import gtsam
import numpy as np

from gtsam import symbol_shorthand

C = symbol_shorthand.C
K = symbol_shorthand.K
P = symbol_shorthand.P
S = symbol_shorthand.S
X = symbol_shorthand.X

Mathematical idea¶

A binary measurement record packages an observation $z_{ab}$ and covariance $\Sigma_{ab}$ with its ordered endpoint keys $(a,b)$ :

(a,b,z_{ab},\Sigma_{ab}).

(1)

It deliberately does not define a prediction $h(x_a,x_b)$ or residual $h(x_a,x_b)-z_{ab}$ ; an algorithm consuming the record chooses that model.

When to use it¶

BinaryMeasurement<T> is data, not an optimizable factor. It keeps an ordered key pair, a value, and its noise model together. The edge direction matters: (a, b) is not interchangeable with (b, a).

Python exposes BinaryMeasurementUnit3, BinaryMeasurementRot3, and BinaryMeasurementPoint3. The wrapper support containers BinaryMeasurementsUnit3, BinaryMeasurementsRot3, and BinaryMeasurementsPoint3 are mainly useful for MATLAB interoperability; ordinary Python code can pass a list.

direction_noise = gtsam.noiseModel.Isotropic.Sigma(2, 0.01)
measurement = gtsam.BinaryMeasurementUnit3(
    X(0), X(1), gtsam.Unit3(np.array([1.0, 0.0, 0.0])), direction_noise
)

print("keys:", measurement.key1(), measurement.key2())
print("direction:", measurement.measured().point3())
print("noise sigmas:", measurement.noiseModel().sigmas())

keys: 8646911284551352320 8646911284551352321
direction: [1. 0. 0.]
noise sigmas: [0.01 0.01]

Practical notes¶

Choose a noise-model dimension matching the manifold dimension of T: 2 for Unit3, 3 for Rot3 and Point3.
Use Unit3 measurements for translation or bearing directions, Rot3 for rotation averaging, and Point3 when magnitude is known.
See TranslationRecovery, GlobalPositioner, and MFAS for algorithms that consume these measurements.

Source¶

BinaryMeasurement.h