w2img::LinearFitEstimator Class Reference

List of all members.

---

Detailed Description

The Kalman linear estimator gives an estimate of the value assuming that the observation is of a linearly changing value (constant velocity for position observations, constant accelation for velocity observations, etc.

)

Unlike the KalmanLinearEstimator, this simply fits a line to the observations and reports the slope.

Public Member Functions
	LinearFitEstimator ()
void	update (float z_k, int MAXIMUM_HISTORY=5)
	update the filter with an observation.
float	getValue () const
	get an estimate of last value
float	getRateOfChange () const
	get the slope of the fitted line.

---

Constructor & Destructor Documentation

w2img::LinearFitEstimator::LinearFitEstimator

(

)

---

Member Function Documentation

float w2img::LinearFitEstimator::getRateOfChange

(

)

const

get the slope of the fitted line.

float w2img::LinearFitEstimator::getValue

(

)

const

get an estimate of last value

void w2img::LinearFitEstimator::update	(	float	z_k,
		int	MAXIMUM_HISTORY = 5
	)

update the filter with an observation.

Specify also how many previous observations should be used. This will automatically prune the list of observations held to that number. The default is to use 5 numbers.

If MAXIMUM_HISTORY is negative, an unlimited history will be used. WARNING: unlike the Kalman estimator, this is very expensive memory-wise since we have to store all the data. Also, unlike the Kalman estimator, there is no way for the moving average to place more weight on later observations, so a positive MAXIMUM_HISTORY is recommended.

---