bayesflt_8hpp-source.html

Bayes滤波器算法C++类说明文档,源码见Bayes滤波器算法

00510 <span class="comment">        Updates x(k|k), X(k|k)</span>
00511 <span class="comment">    */</span>
00512                         
00513     <span class="comment">// Minimum allowable reciprocal condition number for PD Matrix factorisations</span>
<a name="l00514"></a><a class="code" href="classBayesian__filter_1_1Kalman__state__filter.html#o1">00514</a>     <a class="code" href="classBayesian__filter_1_1Numerical__rcond.html">Numerical_rcond</a> <a class="code" href="classBayesian__filter_1_1Kalman__state__filter.html#o1">rclimit</a>;
00515 };
00516 
00517 
00518 <span class="comment">/*</span>
00519 <span class="comment"> * Information State Filter - Abstract filtering property</span>
00520 <span class="comment"> * Linear filter information space representation for 1st (mean) and 2nd (covariance) moments of a distribution</span>
00521 <span class="comment"> *   Y = inv(X)   Information</span>
00522 <span class="comment"> *   y = Y*x      Information state</span>
00523 <span class="comment"> */</span>
00524 
<a name="l00525"></a><a class="code" href="classBayesian__filter_1_1Information__state__filter.html">00525</a> <span class="keyword">class </span><a class="code" href="classBayesian__filter_1_1Information__state__filter.html">Information_state_filter</a> : <span class="keyword">virtual</span> <span class="keyword">public</span> <a class="code" href="classBayesian__filter_1_1Bayes__filter__base.html">Bayes_filter_base</a>
00526 {
00527 <span class="keyword">public</span>:
00528     <a class="code" href="classBayesian__filter_1_1Information__state__filter.html#a0">Information_state_filter</a> (size_t x_size);
<a name="l00529"></a><a class="code" href="classBayesian__filter_1_1Information__state__filter.html#o0">00529</a>     FM::Vec <a class="code" href="classBayesian__filter_1_1Information__state__filter.html#o0">y</a>;              <span class="comment">// Information state</span>
<a name="l00530"></a><a class="code" href="classBayesian__filter_1_1Information__state__filter.html#o1">00530</a>     FM::SymMatrix <a class="code" href="classBayesian__filter_1_1Information__state__filter.html#o1">Y</a>;        <span class="comment">// Information</span>
00531 
00532     <span class="keyword">virtual</span> <span class="keywordtype">void</span> <a class="code" href="classBayesian__filter_1_1Information__state__filter.html#a1">init_yY</a> () =0;
00533     <span class="comment">/* Initialise from a information state and information</span>
00534 <span class="comment">        Requires y(k|k), Y(k|k)</span>
00535 <span class="comment">        Parameters that reference the instance's y and Y members are valid</span>
00536 <span class="comment">    */</span>
00537     <span class="keywordtype">void</span> <a class="code" href="classBayesian__filter_1_1Information__state__filter.html#a2">init_information</a> (<span class="keyword">const</span> FM::Vec&amp; y, <span class="keyword">const</span> FM::SymMatrix&amp; Y);
00538     <span class="comment">/* Initialise from a information state and information</span>
00539 <span class="comment">        Parameters that reference the instance's y and Y members are valid</span>
00540 <span class="comment">    */</span>
00541 
00542     <span class="keyword">virtual</span> <span class="keywordtype">void</span> <a class="code" href="classBayesian__filter_1_1Information__state__filter.html#a3">update_yY</a> () =0;
00543     <span class="comment">/* Update filters information state and information</span>
00544 <span class="comment">        Updates y(k|k), Y(k|k)</span>
00545 <span class="comment">    */</span>
00546 };
00547 
00548 
00549 <span class="comment">/*</span>
00550 <span class="comment"> * Linearizable filter models - Abstract filtering property</span>
00551 <span class="comment"> *  Linrz == A linear, or gradient Linearized filter</span>
00552 <span class="comment"> *</span>
00553 <span class="comment"> * Predict uses a Linrz_predict_model that maintains a Jacobian matrix Fx and addative noise</span>
00554 <span class="comment"> * NOTE: Functional (non-stochastic) predict is NOT possible as predict requires Fx.</span>
00555 <span class="comment"> *</span>
00556 <span class="comment"> * Observe uses a Linrz_observe_model and a variable size observation (z)</span>
00557 <span class="comment"> * There are two variants for correlated and uncorrelated observation noise</span>
00558 <span class="comment"> * Derived filters supply the init,predict,observe,update functions and determine</span>
00559 <span class="comment"> * the algorithm used to implement the filter.</span>
00560 <span class="comment"> */</span>
00561 
<a name="l00562"></a><a class="code" href="classBayesian__filter_1_1Linrz__filter.html">00562</a> <span class="keyword">class </span><a class="code" href="classBayesian__filter_1_1Linrz__filter.html">Linrz_filter</a> : <span class="keyword">virtual</span> <span class="keyword">public</span> <a class="code" href="classBayesian__filter_1_1Bayes__filter__base.html">Bayes_filter_base</a>
00563 { 
00564 <span class="keyword">public</span>:
00565     <span class="comment">/* Virtual functions for filter algorithm */</span>
00566 
00567     <span class="keyword">virtual</span> <a class="code" href="classBayesian__filter_1_1Bayes__base.html#w0">Float</a> <a class="code" href="classBayesian__filter_1_1Linrz__filter.html#a0">predict</a> (<a class="code" href="classBayesian__filter_1_1Linrz__predict__model.html">Linrz_predict_model</a>&amp; f) = 0;
00568     <span class="comment">/* Predict state using a Linrz model</span>
00569 <span class="comment">        Requires x(k|k), X(k|k) or internal equivilent</span>
00570 <span class="comment">        Returns: Reciprocal condition number of primary matrix used in predict computation (1. if none)</span>
00571 <span class="comment">    */</span>
00572 
00573     <span class="keyword">virtual</span> <a class="code" href="classBayesian__filter_1_1Bayes__base.html#w0">Float</a> <a class="code" href="classBayesian__filter_1_1Linrz__filter.html#a1">observe</a> (<a class="code" href="classBayesian__filter_1_1Linrz__uncorrelated__observe__model.html">Linrz_uncorrelated_observe_model</a>&amp; h, <span class="keyword">const</span> FM::Vec&amp; z) = 0;
00574     <span class="keyword">virtual</span> <a class="code" href="classBayesian__filter_1_1Bayes__base.html#w0">Float</a> <a class="code" href="classBayesian__filter_1_1Linrz__filter.html#a1">observe</a> (<a class="code" href="classBayesian__filter_1_1Linrz__correlated__observe__model.html">Linrz_correlated_observe_model</a>&amp; h, <span class="keyword">const</span> FM::Vec&amp; z) = 0;
00575     <span class="comment">/* Observation z(k) and with (Un)correlated observation noise model</span>
00576 <span class="comment">        Requires x(k|k), X(k|k) or internal equivilent</span>
00577 <span class="comment">        Returns: Reciprocal condition number of primary matrix used in observe computation (1. if none)</span>
00578 <span class="comment">    */</span>
00579 };
00580 
00581 
00582 <span class="comment">/*</span>
00583 <span class="comment"> * Linearizable Kalman Filter</span>
00584 <span class="comment"> *  Kalman state representation and linearizable models</span>
00585 <span class="comment"> *</span>
00586 <span class="comment"> * Common abstration for many linear filters</span>
00587 <span class="comment"> *  Has a virtual base to represent the common state</span>
00588 <span class="comment"> */</span>
<a name="l00589"></a><a class="code" href="classBayesian__filter_1_1Linrz__kalman__filter.html">00589</a> <span class="keyword">class </span><a class="code" href="classBayesian__filter_1_1Linrz__kalman__filter.html">Linrz_kalman_filter</a> : <span class="keyword">public</span> <a class="code" href="classBayesian__filter_1_1Linrz__filter.html">Linrz_filter</a>, <span class="keyword">virtual</span> <span class="keyword">public</span> <a class="code" href="classBayesian__filter_1_1Kalman__state__filter.html">Kalman_state_filter</a>
00590 {
00591 <span class="keyword">protected</span>:
<a name="l00592"></a><a class="code" href="classBayesian__filter_1_1Linrz__kalman__filter.html#b0">00592</a>     <a class="code" href="classBayesian__filter_1_1Linrz__kalman__filter.html#b0">Linrz_kalman_filter</a>() : <a class="code" href="classBayesian__filter_1_1Kalman__state__filter.html">Kalman_state_filter</a>(0) <span class="comment">// define a default constructor</span>
00593     {}
00594 };
00595 
00596 
00597 <span class="comment">/*</span>
00598 <span class="comment"> * Extended Kalman Filter</span>
00599 <span class="comment"> *  Kalman state representation and linearizable models</span>
00600 <span class="comment"> *</span>
00601 <span class="comment"> * Observe is implemented using an innovation computed from the non-linear part of the</span>
00602 <span class="comment"> * obseve model and linear part of the Linrz_observe_model</span>
00603 <span class="comment"> *</span>
00604 <span class="comment"> * Common abstration for many linear filters</span>
00605 <span class="comment"> *  Has a virtual base to represent the common state</span>
00606 <span class="comment"> */</span>
<a name="l00607"></a><a class="code" href="classBayesian__filter_1_1Extended__kalman__filter.html">00607</a> <span class="keyword">class </span><a class="code" href="classBayesian__filter_1_1Extended__kalman__filter.html">Extended_kalman_filter</a> : <span class="keyword">public</span> <a class="code" href="classBayesian__filter_1_1Linrz__kalman__filter.html">Linrz_kalman_filter</a>
00608 {
00609 <span class="keyword">protected</span>:
<a name="l00610"></a><a class="code" href="classBayesian__filter_1_1Extended__kalman__filter.html#b0">00610</a>     <a class="code" href="classBayesian__filter_1_1Extended__kalman__filter.html#b0">Extended_kalman_filter</a>() : <a class="code" href="classBayesian__filter_1_1Kalman__state__filter.html">Kalman_state_filter</a>(0) <span class="comment">// define a default constructor</span>
00611     {}
00612 <span class="keyword">public</span>:
00613     <span class="keyword">virtual</span> <a class="code" href="classBayesian__filter_1_1Bayes__base.html#w0">Float</a> <a class="code" href="classBayesian__filter_1_1Extended__kalman__filter.html#a0">observe</a> (Linrz_uncorrelated_observe_model&amp; h, <span class="keyword">const</span> FM::Vec&amp; z);
00614     <span class="keyword">virtual</span> <a class="code" href="classBayesian__filter_1_1Bayes__base.html#w0">Float</a> <a class="code" href="classBayesian__filter_1_1Extended__kalman__filter.html#a0">observe</a> (Linrz_correlated_observe_model&amp; h, <span class="keyword">const</span> FM::Vec&amp; z);
00615     <span class="comment">/* Observation z(k) and with (Un)correlated observation noise model</span>
00616 <span class="comment">        Requires x(k|k), X(k|k) or internal equivilent</span>
00617 <span class="comment">        Returns: Reciprocal condition number of primary matrix used in observe computation (1. if none)</span>
00618 <span class="comment">        Default implementation simple computes innovation for observe_innovation</span>
00619 <span class="comment">    */</span>
00620 
00621     <span class="keyword">virtual</span> <a class="code" href="classBayesian__filter_1_1Bayes__base.html#w0">Float</a> <a class="code" href="classBayesian__filter_1_1Extended__kalman__filter.html#a2">observe_innovation</a> (Linrz_uncorrelated_observe_model&amp; h, <span class="keyword">const</span> FM::Vec&amp; s) = 0;
00622     <span class="keyword">virtual</span> <a class="code" href="classBayesian__filter_1_1Bayes__base.html#w0">Float</a> <a class="code" href="classBayesian__filter_1_1Extended__kalman__filter.html#a2">observe_innovation</a> (Linrz_correlated_observe_model&amp; h, <span class="keyword">const</span> FM::Vec&amp; s) = 0;
00623     <span class="comment">/* Observation innovation s(k) and with (Un)correlated observation noise model</span>
00624 <span class="comment">        Requires x(k|k), X(k|k) or internal equivilent</span>
00625 <span class="comment">        Returns: Reciprocal condition number of primary matrix used in observe computation (1. if none)</span>
00626 <span class="comment">    */</span>
00627 };
00628 
00629 
00630 <span class="comment">/*</span>
00631 <span class="comment"> * Sample State Filter - Abstract filtering property</span>
00632 <span class="comment"> *</span>
00633 <span class="comment"> * Probability distributions are represted by a finite sampling</span>
00634 <span class="comment"> *</span>
00635 <span class="comment"> * State (x_size) size and its sampling (s_size) are assumed to remain constant.</span>
00636 <span class="comment"> * The state sampling public so they can be directly manipulated.</span>
00637 <span class="comment"> *  init: Should be used if they to be altered</span>
00638 <span class="comment"> *  updat