KalmanFilter Class Reference

Simple 1D Kalman Filter for smoothing measurements. More...

#include <KalmanFilter.h>

Public Member Functions
	KalmanFilter (float process_noise=0.01f, float measurement_noise=1.0f)
	Construct a new Kalman Filter object.
void	addMeasurement (float measurement)
	Updates the filter with a new measurement and returns the filtered value.
bool	begin ()
	Reset the filter state to zero, keeping the current noise parameters.
bool	begin (float process_noise, float measurement_noise)
	reset the filter with new parameters
float	calculate ()
	Returns the current estimated value.
void	end ()
	End or clear the filter (sets the estimate to zero).

Protected Attributes
float	K
	Kalman gain (K)
float	P
	Estimation error covariance (P)
float	Q
	Process noise covariance (Q)
float	R
	Measurement noise covariance (R)
float	X
	Estimated state (X)

Detailed Description

Simple 1D Kalman Filter for smoothing measurements.

This class implements a basic one-dimensional Kalman filter for smoothing noisy measurements. It is useful for sensor data fusion and signal processing applications where noise reduction is required.

The Kalman filter uses two main parameters:

• process_noise (Q): Represents the expected process variance. A typical default is 0.01.
• measurement_noise (R): Represents the expected measurement variance. A typical default is 1.0.

Lower values for Q make the filter trust the model more (less responsive to changes), while higher values for R make the filter trust the measurements less (more smoothing).

Note

Reasonable defaults for many sensor applications are Q = 0.01 and R = 1.0, but these should be tuned for your specific use case.

Author

Phil Schatzmann

Constructor & Destructor Documentation

KalmanFilter()

KalmanFilter ( float  process_noise = 0.01f, float  measurement_noise = 1.0f  )

inline

Construct a new Kalman Filter object.

Parameters

process_noise	The process noise covariance (Q). Default is 0.01.
measurement_noise	The measurement noise covariance (R). Default is 1.0.

Member Function Documentation

addMeasurement()

void addMeasurement ( float  measurement )

inline

Updates the filter with a new measurement and returns the filtered value.

Parameters

measurement

The new measurement to be filtered

Returns

float The updated (filtered) estimate

begin() [1/2]

bool begin ( )

inline

Reset the filter state to zero, keeping the current noise parameters.

Returns

true Always returns true.

begin() [2/2]

bool begin ( float  process_noise, float  measurement_noise  )

inline

reset the filter with new parameters

Parameters

process_noise	The process noise covariance (Q). Default is 0.01.
measurement_noise	The measurement noise covariance (R). Default is 1.0.

Initialize or reset the filter with new parameters.

Parameters

process_noise	The process noise covariance (Q). Default is 0.01.
measurement_noise	The measurement noise covariance (R). Default is 1.0.

Returns

true Always returns true.

calculate()

float calculate ( )

inline

Returns the current estimated value.

Returns

float The current estimate

end()

void end ( )

inline

End or clear the filter (sets the estimate to zero).

Member Data Documentation

K

float K

protected

Kalman gain (K)

P

float P

protected

Estimation error covariance (P)

Q

float Q

protected

Process noise covariance (Q)

R

float R

protected

Measurement noise covariance (R)

X

float X

protected

Estimated state (X)

---

The documentation for this class was generated from the following file:

• src/AudioTools/CoreAudio/AudioBasic/KalmanFilter.h