Struct RandomMeasurementMatrix 

pub struct RandomMeasurementMatrix {
    pub m: usize,
    pub n: usize,
    pub matrix: Vec<Vec<f64>>,
}

A random Gaussian measurement matrix for compressed sensing.

Each row is an independent Gaussian random vector; m << n enables sub-Nyquist recovery of sparse signals.

Fields

m: usize

Number of measurements (rows).

n: usize

Signal length (columns).

matrix: Vec<Vec<f64>>

Underlying matrix entries stored row-major.

Implementations

impl RandomMeasurementMatrix

pub fn generate_gaussian(m: usize, n: usize) -> Self

Generate an m × n Gaussian measurement matrix (entries ~ N(0, 1/m)).

The columns are scaled by 1/sqrt(m) so that each measurement approximately preserves the signal energy.

pub fn generate_bernoulli(m: usize, n: usize) -> Self

Generate an m × n Bernoulli ±1/sqrt(m) measurement matrix.

Each entry is independently ±1/sqrt(m) with equal probability 1/2.

pub fn measure(&self, x: &[f64]) -> Vec<f64>

Apply the measurement matrix to signal x.

Returns a vector of length m (the compressed measurements).

pub fn coherence(&self) -> f64

Compute the mutual coherence of the measurement matrix columns.

Lower coherence is better for sparse recovery.