Fix a bug in getb and replace it with negated values of gradientAtZero. Add some comments about a bug.
parent
f9d6c3da22
commit
60f43c7a4b
|
@ -3,6 +3,7 @@
|
||||||
*
|
*
|
||||||
* Created on: Feb 14, 2012
|
* Created on: Feb 14, 2012
|
||||||
* Author: ydjian
|
* Author: ydjian
|
||||||
|
* Author: Sungtae An
|
||||||
*/
|
*/
|
||||||
|
|
||||||
#include <gtsam/linear/GaussianFactorGraph.h>
|
#include <gtsam/linear/GaussianFactorGraph.h>
|
||||||
|
@ -80,9 +81,13 @@ void GaussianFactorGraphSystem::residual(const Vector &x, Vector &r) const {
|
||||||
void GaussianFactorGraphSystem::multiply(const Vector &x, Vector& AtAx) const {
|
void GaussianFactorGraphSystem::multiply(const Vector &x, Vector& AtAx) const {
|
||||||
/* implement A^t*A*x, assume x and AtAx are pre-allocated */
|
/* implement A^t*A*x, assume x and AtAx are pre-allocated */
|
||||||
|
|
||||||
|
// Build a VectorValues for Vector x
|
||||||
VectorValues vvX = buildVectorValues(x,keyInfo_);
|
VectorValues vvX = buildVectorValues(x,keyInfo_);
|
||||||
|
// VectorValues form of A'Ax for multiplyHessianAdd
|
||||||
VectorValues vvAtAx;
|
VectorValues vvAtAx;
|
||||||
|
// vvAtAx += 1.0 * A'Ax for each factor
|
||||||
gfg_.multiplyHessianAdd(1.0, vvX, vvAtAx);
|
gfg_.multiplyHessianAdd(1.0, vvX, vvAtAx);
|
||||||
|
// Make the result as Vector form
|
||||||
AtAx = vvAtAx.vector();
|
AtAx = vvAtAx.vector();
|
||||||
|
|
||||||
}
|
}
|
||||||
|
@ -91,30 +96,46 @@ void GaussianFactorGraphSystem::multiply(const Vector &x, Vector& AtAx) const {
|
||||||
void GaussianFactorGraphSystem::getb(Vector &b) const {
|
void GaussianFactorGraphSystem::getb(Vector &b) const {
|
||||||
/* compute rhs, assume b pre-allocated */
|
/* compute rhs, assume b pre-allocated */
|
||||||
|
|
||||||
/* reset */
|
/* ------------------------------------------------------------------------
|
||||||
b.setZero();
|
* Multiply and getb functions (build function in preconditioner.cpp also)
|
||||||
|
* Yong-Dian's code had a bug that they do not consider noise model
|
||||||
|
* which means that they do not whiten A and b.
|
||||||
|
* It has no problem when the associated noise model has a form of Isotropic
|
||||||
|
* because it can be cancelled out on both l.h.s and r.h.s of equation.
|
||||||
|
* However, it cause a wrong result with non-isotropic noise model.
|
||||||
|
* The unit test for PCSSolver (testPCGSolver.cpp) Yond-Dian made use a
|
||||||
|
* example factor graph which has isotropic noise model and
|
||||||
|
* that is the reason why there was no unit test error.
|
||||||
|
* ------------------------------------------------------------------------*/
|
||||||
|
// /* reset */
|
||||||
|
// b.setZero();
|
||||||
|
//
|
||||||
|
// BOOST_FOREACH ( const GaussianFactor::shared_ptr &gf, gfg_ ) {
|
||||||
|
// if ( JacobianFactor::shared_ptr jf = boost::dynamic_pointer_cast<JacobianFactor>(gf) ) {
|
||||||
|
// const Vector rhs = jf->getb();
|
||||||
|
// /* accumulate At rhs */
|
||||||
|
// for ( JacobianFactor::const_iterator it = jf->begin() ; it != jf->end() ; ++it ) {
|
||||||
|
// /* this map lookup should be replaced */
|
||||||
|
// const KeyInfoEntry &entry = keyInfo_.find(*it)->second;
|
||||||
|
// b.segment(entry.colstart(), entry.dim()) += jf->getA(it).transpose() * rhs ;
|
||||||
|
// }
|
||||||
|
// }
|
||||||
|
// else if ( HessianFactor::shared_ptr hf = boost::dynamic_pointer_cast<HessianFactor>(gf) ) {
|
||||||
|
// /* accumulate g */
|
||||||
|
// for (HessianFactor::const_iterator it = hf->begin(); it != hf->end(); it++) {
|
||||||
|
// const KeyInfoEntry &entry = keyInfo_.find(*it)->second;
|
||||||
|
// b.segment(entry.colstart(), entry.dim()) += hf->linearTerm(it);
|
||||||
|
// }
|
||||||
|
// }
|
||||||
|
// else {
|
||||||
|
// throw invalid_argument("GaussianFactorGraphSystem::getb gfg contains a factor that is neither a JacobianFactor nor a HessianFactor.");
|
||||||
|
// }
|
||||||
|
// }
|
||||||
|
|
||||||
BOOST_FOREACH ( const GaussianFactor::shared_ptr &gf, gfg_ ) {
|
// Get whitened r.h.s (b vector) from each factor in the form of VectorValues
|
||||||
if ( JacobianFactor::shared_ptr jf = boost::dynamic_pointer_cast<JacobianFactor>(gf) ) {
|
VectorValues vvb = gfg_.gradientAtZero();
|
||||||
const Vector rhs = jf->getb();
|
// Make the result as Vector form
|
||||||
/* accumulate At rhs */
|
b = -vvb.vector();
|
||||||
for ( JacobianFactor::const_iterator it = jf->begin() ; it != jf->end() ; ++it ) {
|
|
||||||
/* this map lookup should be replaced */
|
|
||||||
const KeyInfoEntry &entry = keyInfo_.find(*it)->second;
|
|
||||||
b.segment(entry.colstart(), entry.dim()) += jf->getA(it).transpose() * rhs ;
|
|
||||||
}
|
|
||||||
}
|
|
||||||
else if ( HessianFactor::shared_ptr hf = boost::dynamic_pointer_cast<HessianFactor>(gf) ) {
|
|
||||||
/* accumulate g */
|
|
||||||
for (HessianFactor::const_iterator it = hf->begin(); it != hf->end(); it++) {
|
|
||||||
const KeyInfoEntry &entry = keyInfo_.find(*it)->second;
|
|
||||||
b.segment(entry.colstart(), entry.dim()) += hf->linearTerm(it);
|
|
||||||
}
|
|
||||||
}
|
|
||||||
else {
|
|
||||||
throw invalid_argument("GaussianFactorGraphSystem::getb gfg contains a factor that is neither a JacobianFactor nor a HessianFactor.");
|
|
||||||
}
|
|
||||||
}
|
|
||||||
}
|
}
|
||||||
|
|
||||||
/**********************************************************************************/
|
/**********************************************************************************/
|
||||||
|
|
Loading…
Reference in New Issue