This no_std
Rust crate implements a dynamic smoothing filter based on an algorithm by
Andrew Simper.
It uses a lowpass filter whose cutoff frequency is modulated dynamically by the signal itself and allows to track signal changes faster than a purely static version.
The main use of this crate is to smooth noisy or stepped signals like analog measurements from an ADC in embedded systems.
Currently there are only implementations of the "efficient" variant using single-precision floats and 32-bit integers.
use dyn_smooth::DynamicSmootherEcoF32;
// Create an instance with suitable settings.
let base_freq = 2.0;
let sample_freq = 1000.0;
let sensitivity = 0.5;
let mut smoother = DynamicSmootherEcoF32::new(base_freq, sample_freq, sensitivity);
// Feed an input value to the smoother and retrieve the smoothed value.
for sample in 0..100 {
let input_value = sample as f32; // Dummy value, read some real value from somewhere
let smoothed_value = smoother.tick(input_value);
}
use dyn_smooth::{DynamicSmootherEcoI32, I32_FRAC_BITS};
// Create an instance with suitable settings.
let base_freq = 2 << I32_FRAC_BITS;
let sample_freq = 1000 << I32_FRAC_BITS;
let sensitivity = (0.5 * ((1 << I32_FRAC_BITS) as f32)) as i32;
let mut smoother = DynamicSmootherEcoI32::new(base_freq, sample_freq, sensitivity);
// Feed an input value to the smoother and retrieve the smoothed value.
for sample in 0..100 {
let input_value = sample as i32; // Dummy value, read some real value from somewhere
let smoothed_value = smoother.tick(input_value);
}
Run cargo test
for the unit tests. Use the --nocapture
option for additional output.
Published under the MIT license. All contributions to this project must be provided under the same license conditions.
Author: Oliver Rockstedt info@sourcebox.de