types.core.CovarianceMatrix

types.core.CovarianceMatrix

Covariance matrix (symmetric, positive semidefinite).

Represents uncertainty or cost weighting.

Uses: - State uncertainty: P (nx, nx) - Measurement noise: R (ny, ny) - Process noise: Q (nx, nx) or (nw, nw) - Parameter uncertainty: Σ_θ (np, np)

Always symmetric: Σ = Σ’ Always PSD: v’Σv ≥ 0 for all v

Examples

>>> # State covariance (isotropic)
>>> P: CovarianceMatrix = np.eye(3)
>>> 
>>> # Measurement noise (diagonal)
>>> R: CovarianceMatrix = np.diag([0.1, 0.05, 0.02])
>>> 
>>> # Process noise (correlated)
>>> Q: CovarianceMatrix = np.array([[0.1, 0.05], [0.05, 0.2]])