Large gtsam refactoring
To support faster development *and* better performance Richard and I pushed through a large refactoring of NonlinearFactors.
The following are the biggest changes:
1) NonLinearFactor1 and NonLinearFactor2 are now templated on Config, Key type, and X type, where X is the argument to the measurement function.
2) The measurement itself is no longer kept in the nonlinear factor. Instead, a derived class (see testVSLAMFactor, testNonlinearEquality, testPose3Factor etc...) has to implement a function to compute the errors, "evaluateErrors". Instead of (h(x)-z), it needs to return (z-h(x)), so Ax-b is an approximation of the error. IMPORTANT: evaluateErrors needs - if asked - *combine* the calculation of the function value h(x) and the derivatives dh(x)/dx. This was a major performance issue. To do this, boost::optional<Matrix&> arguments are provided, and tin EvaluateErrors you just says something like
if (H) *H = Matrix_(3,6,....);
3) We are no longer using int or strings for nonlinear factors. Instead, the preferred key type is now Symbol, defined in Key.h. This is both fast and cool: you can construct it from an int, and cast it to a strong. It also does type checking: a Symbol<Pose3,'x'> will not match a Symbol<Pose2,'x'>
4) minor: take a look at LieConfig.h: it help you avoid writing a lot of code bu automatically creating configs for a certain type. See e.g. Pose3Config.h. A "double" LieConfig is on the way - Thanks Richard and Manohar !
2010-01-14 06:25:03 +08:00
|
|
|
/*
|
|
|
|
|
* graph.h
|
|
|
|
|
* @brief Graph algorithm using boost library
|
|
|
|
|
* @author: Kai Ni
|
|
|
|
|
* Created on: Jan 11, 2010
|
|
|
|
|
*/
|
|
|
|
|
|
|
|
|
|
#pragma once
|
|
|
|
|
|
|
|
|
|
#include <map>
|
|
|
|
|
|
|
|
|
|
#include <boost/graph/graph_traits.hpp>
|
|
|
|
|
#include <boost/graph/adjacency_list.hpp>
|
|
|
|
|
#include <boost/shared_ptr.hpp>
|
|
|
|
|
|
|
|
|
|
namespace gtsam {
|
|
|
|
|
|
|
|
|
|
// type definitions :
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* SDGraph is undirected graph with variable keys and double edge weights
|
|
|
|
|
*/
|
|
|
|
|
template<class Key>
|
|
|
|
|
class SDGraph: public boost::adjacency_list<boost::vecS, boost::vecS, boost::undirectedS,
|
|
|
|
|
boost::property<boost::vertex_name_t, Key>, boost::property<
|
|
|
|
|
boost::edge_weight_t, double> > {
|
2010-01-14 11:21:07 +08:00
|
|
|
public:
|
|
|
|
|
typedef typename boost::graph_traits<SDGraph<Key> >::vertex_descriptor Vertex;
|
Large gtsam refactoring
To support faster development *and* better performance Richard and I pushed through a large refactoring of NonlinearFactors.
The following are the biggest changes:
1) NonLinearFactor1 and NonLinearFactor2 are now templated on Config, Key type, and X type, where X is the argument to the measurement function.
2) The measurement itself is no longer kept in the nonlinear factor. Instead, a derived class (see testVSLAMFactor, testNonlinearEquality, testPose3Factor etc...) has to implement a function to compute the errors, "evaluateErrors". Instead of (h(x)-z), it needs to return (z-h(x)), so Ax-b is an approximation of the error. IMPORTANT: evaluateErrors needs - if asked - *combine* the calculation of the function value h(x) and the derivatives dh(x)/dx. This was a major performance issue. To do this, boost::optional<Matrix&> arguments are provided, and tin EvaluateErrors you just says something like
if (H) *H = Matrix_(3,6,....);
3) We are no longer using int or strings for nonlinear factors. Instead, the preferred key type is now Symbol, defined in Key.h. This is both fast and cool: you can construct it from an int, and cast it to a strong. It also does type checking: a Symbol<Pose3,'x'> will not match a Symbol<Pose2,'x'>
4) minor: take a look at LieConfig.h: it help you avoid writing a lot of code bu automatically creating configs for a certain type. See e.g. Pose3Config.h. A "double" LieConfig is on the way - Thanks Richard and Manohar !
2010-01-14 06:25:03 +08:00
|
|
|
};
|
|
|
|
|
|
2010-01-14 09:25:40 +08:00
|
|
|
template<class Key>
|
|
|
|
|
class SGraph : public boost::adjacency_list<boost::vecS, boost::vecS, boost::directedS,
|
2010-01-14 11:21:07 +08:00
|
|
|
boost::property<boost::vertex_name_t, Key> > {
|
2010-01-14 10:07:14 +08:00
|
|
|
public:
|
2010-01-14 11:21:07 +08:00
|
|
|
typedef typename boost::graph_traits<SGraph<Key> >::vertex_descriptor Vertex;
|
2010-01-14 09:25:40 +08:00
|
|
|
};
|
|
|
|
|
|
|
|
|
|
//typedef boost::graph_traits<SGraph>::vertex_descriptor SVertex;
|
|
|
|
|
|
Large gtsam refactoring
To support faster development *and* better performance Richard and I pushed through a large refactoring of NonlinearFactors.
The following are the biggest changes:
1) NonLinearFactor1 and NonLinearFactor2 are now templated on Config, Key type, and X type, where X is the argument to the measurement function.
2) The measurement itself is no longer kept in the nonlinear factor. Instead, a derived class (see testVSLAMFactor, testNonlinearEquality, testPose3Factor etc...) has to implement a function to compute the errors, "evaluateErrors". Instead of (h(x)-z), it needs to return (z-h(x)), so Ax-b is an approximation of the error. IMPORTANT: evaluateErrors needs - if asked - *combine* the calculation of the function value h(x) and the derivatives dh(x)/dx. This was a major performance issue. To do this, boost::optional<Matrix&> arguments are provided, and tin EvaluateErrors you just says something like
if (H) *H = Matrix_(3,6,....);
3) We are no longer using int or strings for nonlinear factors. Instead, the preferred key type is now Symbol, defined in Key.h. This is both fast and cool: you can construct it from an int, and cast it to a strong. It also does type checking: a Symbol<Pose3,'x'> will not match a Symbol<Pose2,'x'>
4) minor: take a look at LieConfig.h: it help you avoid writing a lot of code bu automatically creating configs for a certain type. See e.g. Pose3Config.h. A "double" LieConfig is on the way - Thanks Richard and Manohar !
2010-01-14 06:25:03 +08:00
|
|
|
/**
|
|
|
|
|
* Map from variable key to parent key
|
|
|
|
|
*/
|
|
|
|
|
template <class Key>
|
|
|
|
|
class PredecessorMap : public std::map<Key,Key> {};
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* Convert the factor graph to an SDGraph
|
|
|
|
|
* G = Graph type
|
|
|
|
|
* F = Factor type
|
|
|
|
|
* Key = Key type
|
|
|
|
|
*/
|
|
|
|
|
template<class G, class F, class Key> SDGraph<Key> toBoostGraph(const G& graph);
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* Build takes a predecessor map, and builds a directed graph corresponding to the tree.
|
|
|
|
|
* G = Graph type
|
|
|
|
|
* V = Vertex type
|
|
|
|
|
*/
|
2010-01-14 11:21:07 +08:00
|
|
|
template<class G, class V, class Key>
|
|
|
|
|
boost::tuple<G, V, std::map<Key,V> > predecessorMap2Graph(const PredecessorMap<Key>& p_map);
|
Large gtsam refactoring
To support faster development *and* better performance Richard and I pushed through a large refactoring of NonlinearFactors.
The following are the biggest changes:
1) NonLinearFactor1 and NonLinearFactor2 are now templated on Config, Key type, and X type, where X is the argument to the measurement function.
2) The measurement itself is no longer kept in the nonlinear factor. Instead, a derived class (see testVSLAMFactor, testNonlinearEquality, testPose3Factor etc...) has to implement a function to compute the errors, "evaluateErrors". Instead of (h(x)-z), it needs to return (z-h(x)), so Ax-b is an approximation of the error. IMPORTANT: evaluateErrors needs - if asked - *combine* the calculation of the function value h(x) and the derivatives dh(x)/dx. This was a major performance issue. To do this, boost::optional<Matrix&> arguments are provided, and tin EvaluateErrors you just says something like
if (H) *H = Matrix_(3,6,....);
3) We are no longer using int or strings for nonlinear factors. Instead, the preferred key type is now Symbol, defined in Key.h. This is both fast and cool: you can construct it from an int, and cast it to a strong. It also does type checking: a Symbol<Pose3,'x'> will not match a Symbol<Pose2,'x'>
4) minor: take a look at LieConfig.h: it help you avoid writing a lot of code bu automatically creating configs for a certain type. See e.g. Pose3Config.h. A "double" LieConfig is on the way - Thanks Richard and Manohar !
2010-01-14 06:25:03 +08:00
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* Compose the poses by following the chain specified by the spanning tree
|
|
|
|
|
*/
|
|
|
|
|
template<class G, class Factor, class Pose, class Config>
|
|
|
|
|
boost::shared_ptr<Config>
|
|
|
|
|
composePoses(const G& graph, const PredecessorMap<typename Config::Key>& tree, const Pose& rootPose);
|
|
|
|
|
|
|
|
|
|
} // namespace gtsam
|
