average1.hpp

Go to the documentation of this file.

#ifndef _BAYES_FILTER_AVERAGE1
#define _BAYES_FILTER_AVERAGE1
 
/*
 * Bayes++ the Bayesian Filtering Library
 * Copyright (c) 2002 Michael Stevens
 * See accompanying Bayes++.htm for terms and conditions of use.
 *
 * $Id$
 */
 
/*
 * Predefined filter: Average1_filter
 *  A single state averager
 */
 
/* Filter namespace */
namespace Bayesian_filter
{
 
template <class Filter_base>
class Average1_filter
{
    Filter_base ksf;        // Kalman State Filter
    
    typedef typename Filter_base::Float Float;
    class Cpredict : public Linear_predict_model
    // Constant predict model
    {
    public:
        Cpredict(Float qq) : Linear_predict_model(1, 1)
        {
            Fx(0,0) = 1.;
            q[0] = qq;
            G(0,0) = 1.;
        }
    };
 
    class Cobserve : public Linear_correlated_observe_model
    // Constant observe model
    {
    public:
        Cobserve(Float ZZ) : Linear_correlated_observe_model(1,1)
        {
            Hx(0,0) = 1.;
            Z(0,0) = ZZ;
        }
    };
 
public:
    Average1_filter (Float iQ, Float iZ, Float z);
    Average1_filter (Float iQ, Float iZ);
    Float observe (Float zz);
    operator Float () const
    /* Returns filtered estimate
     */
    {   if (!bInit)
            ksf.error (Logic_exception("Average1 not init"));
        return ksf.x[0];
    }
 
private:
    Cpredict f;
    Cobserve h;
    
    bool bInit;
    FM::Vec z;
};
 
 
 
template <typename Filter_base>
Average1_filter<Filter_base>::Average1_filter (Float iQ, Float iZ, Float zz) :
    ksf(1), f(iQ), h(iZ), z(1)
/* Initialise noises and set sizes
 * include first observation zz */
{
    bInit = false;
    
    observe (zz);
}
 
template <typename Filter_base>
Average1_filter<Filter_base>::Average1_filter (Float iQ, Float iZ) :
    ksf(1), f(iQ), h(iZ), z(1)
// Initialise noises and set sizes
{
    bInit = false;
}
 
template <typename Filter_base>
typename Average1_filter<Filter_base>::Float Average1_filter<Filter_base>::observe(Float zz)
/* Observe z, first call set initial state to z
 * Returns filtered estimate
 */
{
    z[0] = zz;
 
    if (!bInit) {
        ksf.init_kalman (z, h.Z);
        bInit = true;
    }
 
    ksf.predict(f);
    ksf.observe(h, z);
    ksf.update ();
 
    return ksf.x[0];
}
 
}//namespace
#endif