oRTP 5.0.0
|
#include <utils.h>
Data Fields | |
double | m |
double | b |
double | P [2][2] |
double | lambda |
Utility object to interpolate linear model based on captures with 0-mean noise. Based on the model (x, y) where y evolves depending on x with relation y = m*x+b, it will estimate unknown values b, m using given noisy measures xmes and ymes, eg real system is evolving with: y = m * x + b + noise. Note: If noise is NOT white, the residue will be absorbed by one of the estimators. It is an implementation of recursive least square algorithm, based on Kalman filter.
double _OrtpKalmanRLS::lambda |
Forgetting factor in [94, .999]. Used when unknown parameters vary in time.