Struct Conditioning 

pub struct Conditioning { /* private fields */ }

Conditioning of the problem.

Implementations

impl Conditioning

pub fn centre(&self) -> Vector3<f64>

The centre of rotation the report refers to.

pub fn radius_of_gyration(&self) -> f64

The radius of gyration: the characteristic length of the normalisation.

pub fn used(&self) -> usize

How many correspondences took part.

pub fn singular_values(&self) -> [f64; 6]

The normalised singular values, in decreasing order.

pub fn condition_number(&self) -> f64

The condition number σ_max / σ_min.

Invariant to the scale of the scene and to the choice of units, which is exactly what the normalisation is for.

pub fn direction(&self, index: usize) -> Vector6<f64>

Direction number index in the normalised coordinates ξ'.

pub fn direction_in_world(&self, index: usize) -> Vector6<f64>

The same direction in the original coordinates ξ = [ρ; φ], referred to the coordinate origin.

Undoes the normalisation (φ is divided by r_g) and moves the centre of rotation from the centroid back to the origin through the adjoint of that translation.

pub fn to_normalised(&self, world: Vector6<f64>) -> Vector6<f64>

Converts a pose increment from world coordinates ξ = [ρ; φ] into the normalised ξ' the spectrum is expressed in.

The inverse of direction_in_world: first move the centre of rotation to the centroid, then scale the rotational part by the radius of gyration. Needed to compare a predicted spread against an empirical one — they must live in the same coordinates.

pub fn component(&self, index: usize, world: Vector6<f64>) -> f64

Projection of a pose increment onto direction index, metres.

pub fn uncertainty(&self, noise_sigma: f64) -> [f64; 6]

Standard deviation along each direction, metres.

σ_noise / σ'ᵢ. All six values share units: one unit of a normalised coordinate is one metre of displacement, for rotations too, at distance r_g from the centre.

The point count is accounted for automatically: σ'ᵢ grows as the square root of the sum of weights, so the spread falls as 1/√N.

pub fn classify(&self, criteria: &ObservabilityCriteria) -> [Observability; 6]

Classification of the six degrees of freedom.

pub fn unobservable_directions( &self, criteria: &ObservabilityCriteria, ) -> Vec<Vector6<f64>>

Directions judged unobservable, in ξ coordinates relative to the origin.

Trait Implementations

impl Clone for Conditioning

fn clone(&self) -> Conditioning

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for Conditioning

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.