gtsam  3.2.1
gtsam
 All Classes Namespaces Files Functions Variables Typedefs Enumerations Enumerator Friends Macros Groups Pages
List of all members | Public Types | Protected Member Functions | Protected Attributes
gtsam::ExtendedKalmanFilter< VALUE > Class Template Reference

Detailed Description

template<class VALUE>
class gtsam::ExtendedKalmanFilter< VALUE >

This is a generic Extended Kalman Filter class implemented using nonlinear factors.

GTSAM basically does SRIF with Cholesky to solve the filter problem, making this an efficient, numerically stable Kalman Filter implementation.

The Kalman Filter relies on two models: a motion model that predicts the next state using the current state, and a measurement model that predicts the measurement value at a given state. Because these two models are situation-dependent, base classes for each have been provided above, from which the user must derive a class and incorporate the actual modeling equations.

The class provides a "predict" and "update" function to perform these steps independently. TODO: a "predictAndUpdate" that combines both steps for some computational savings.

Standard Constructors

 ExtendedKalmanFilter (Key key_initial, T x_initial, noiseModel::Gaussian::shared_ptr P_initial)
 

Testable

void print (const std::string &s="") const
 print
 

Advanced Interface

T predict (const MotionFactor &motionFactor)
 TODO: comment.
 
T update (const MeasurementFactor &measurementFactor)
 TODO: comment.
 

Public Types

typedef boost::shared_ptr
< ExtendedKalmanFilter< VALUE > > 		shared_ptr
 
typedef VALUE T
 
typedef NoiseModelFactor2
< VALUE, VALUE > 		MotionFactor
 
typedef NoiseModelFactor1< VALUE > MeasurementFactor
 

Protected Member Functions

T solve_ (const GaussianFactorGraph &linearFactorGraph, const Values &linearizationPoints, Key x, JacobianFactor::shared_ptr &newPrior) const
 

Protected Attributes

T x_
 
JacobianFactor::shared_ptr priorFactor_
 
The documentation for this class was generated from the following files: