Struct hifitime::ClockNoise

pub struct ClockNoise { /* fields omitted */ }

ClockNoise adds true clock drift to a given Duration measurement. For example, if a vehicle is measuring the time of flight of a signal with high precision oscillator, the engineering specifications will include the oscillator stability. This specification bounds the preciseness of time span calculations. On very short time spans, i.e. less than a few minutes, clock drift is usually negligible. However, in several high fidelity systems the clock drift may lead to a significant error (e.g. several kilometers in two-way radar ranging). This module allows high fidelity simulation systems to test the resilience of algorithms with oscillator stability. The constructors here are specified in parts per million: for a parts per billion specification simply multiply the value by 1e-3. NOTE: Clock stability is not linear. If a clock is rated at stable within 15 ppm per fifteen minute interval this does not correspond to 1 ppm per minute.

Example

use hifitime::ClockNoise;

// The IRIS clock is 1 part per billion over one second
let nasa_iris = ClockNoise::with_ppm_over_1sec(1e-3);
let ddoor = 8.0 * 60.0;
let noisy = nasa_iris.noise_up(ddoor);
assert!(
    (noisy - ddoor).abs() < 1e-3,
    "Expected a zero deviation for IRIS"
);

Implementations

impl ClockNoise

pub fn with_ppm(ppm: f64) -> ClockNoise

Creates a new ClockNoise generator from the stability characteristics in absolute parts per million The ppm value is assumed to be the 7-sigma deviation.

pub fn with_ppm_over_1sec(ppm: f64) -> ClockNoise

Creates a new ClockNoise generator from the stability characteristics in parts per million over one second.

pub fn with_ppm_over_1min(ppm: f64) -> ClockNoise

Creates a new ClockNoise generator from the stability characteristics in parts per million over one minute (i.e. 60 seconds).

pub fn with_ppm_over_15min(ppm: f64) -> ClockNoise

Creates a new ClockNoise generator from the stability characteristics in parts per million over fifteen minutes (i.e. 900 seconds).

pub fn noise_up(&self, duration_in_secs: f64) -> f64

From an input set of seconds, returns a random walk number of seconds corresponding to the value plus/minus a drift This is the most accurate method to generate a noisy signal, but it’s extremely slow.

pub fn sample(&self, value: f64) -> f64

Sample the clock for a specific value. Can be used to determined a sampled frequency from an input frequency in Hertz