plBayesianFilter Class Reference

This class implements general purpose Bayesian Filters.
Four types of filters are allowed:
- Filters with action variables and non Idendity system model P(S S_ O | A) = P(S_) P(S | S_ A) P(O | S A).
- Filters with no action variables and non Idendity system model P(S S_ O) = P(S_) P(S | S_) P(O | S).
- Filters with action variables and Idendity system model (i.e P(S | S_) = Dirac(S, S_)) P(S O | A) = P(S) P(O | S A).
- Filters with no action variables and Idendity system model P(S O) = P(S) P(O | S). More...

#include <plBayesianFilter.h>

Inheritance diagram for plBayesianFilter:

[legend]Collaboration diagram for plBayesianFilter:

[legend]List of all members.

Public Member Functions
	plBayesianFilter (const plKernel &InitialState, const plCndKernel &SystemModel, const plCndKernel &ObservationModel, const plVariablesConjunction &action_vars)
	Constuction of a filter with action variables and non Idendity system model: P(S S_ O | A) = P(S_) P(S | S_ A) P(O | S A).
	plBayesianFilter (const plKernel &InitialState, const plVariablesConjunction &old_state_vars, const plCndKernel &SystemModel, const plVariablesConjunction &state_vars, const plCndKernel &ObservationModel, const plVariablesConjunction &obs_vars, const plVariablesConjunction &action_vars)
	Same as above, but with explicitly giving the variables.
	plBayesianFilter (const plKernel &InitialState, const plCndKernel &SystemModel, const plCndKernel &ObservationModel)
	Constuction of a filter with no action variables and non Idendity system model: P(S S_ O) = P(S_) P(S | S_) P(O | S).
	plBayesianFilter (const plKernel &InitialState, const plVariablesConjunction &old_state_vars, const plCndKernel &SystemModel, const plVariablesConjunction &state_vars, const plCndKernel &ObservationModel, const plVariablesConjunction &obs_vars)
	Same as above, but with explicitly giving the variables.
	plBayesianFilter (const plKernel &InitialState, const plCndKernel &ObservationModel, const plVariablesConjunction &action_vars)
	Constuction of a filter with action variables and Identity system model: P(S O | A) = P(S) P(O | S A).
	plBayesianFilter (const plKernel &InitialState, const plCndKernel &ObservationModel)
	Constuction of a filter with no action variables and Identity system model: P(S O) = P(S) P(O | S).
	~plBayesianFilter ()
	Destructor.
void	setAction (const plValues &act)
	Apply new command.
void	prediction ()
	Exact prediction step P(S | A = a) (or P(S) if no action variables).
void	n_prediction (unsigned long int nsamples)
	Prediction step P(S | A = a) (or P(S) if no action variables) using 'nsamples' sample points.
void	time_prediction (double time_in_sec)
	Prediction step P(S | A = a) (or P(S) if no action variables) using for 'time_in_sec' seconds.
void	observation (const plValues &obs)
	Incorporate new observation and do the update step P(S | O = o A = a) (or P(S | O = o) if no action variables).
void	addMeasurement (const plValues &act, const plValues &obs)
	Fix the action, incorporate an observation, and apply prediction/update steps.
void	n_addMeasurement (unsigned long int nsamples, const plValues &act, const plValues &obs)
	Fix the action, compute a prediction using 'nsamples' samples, and incorporate an observation.
void	time_addMeasurement (double time_in_sec, const plValues &act, const plValues &obs)
	Fix the action, compute a prediction in 'time_in_sec' seconds, and incorporate an observation.
void	addMeasurement (const plValues &obs)
	Incorporate an observation, and apply prediction/update steps.
void	n_addMeasurement (unsigned long int nsamples, const plValues &obs)
	Compute a prediction using 'nsamples' samples, and incorporate an observation.
void	time_addMeasurement (double time_in_sec, const plValues &obs)
	Compute a prediction in 'time_in_sec' seconds, and incorporate an observation.
plKernel	getEstimation () const
	Returns the current estimated distribution P(S).

---

Detailed Description

This class implements general purpose Bayesian Filters.
Four types of filters are allowed:
- Filters with action variables and non Idendity system model P(S S_ O | A) = P(S_) P(S | S_ A) P(O | S A).
- Filters with no action variables and non Idendity system model P(S S_ O) = P(S_) P(S | S_) P(O | S).
- Filters with action variables and Idendity system model (i.e P(S | S_) = Dirac(S, S_)) P(S O | A) = P(S) P(O | S A).
- Filters with no action variables and Idendity system model P(S O) = P(S) P(O | S).

Definition at line 42 of file plBayesianFilter.h.

---

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

• plBayesianFilter.h

---