From 168ddf5457ce57377d49d6aabb42f336f27ca39e Mon Sep 17 00:00:00 2001 From: Yong-Dian Jian Date: Tue, 24 Jul 2012 21:06:33 +0000 Subject: [PATCH] add Cal3DS2.calibrate() with fixed point iteration reorg nonlinear conjugate gradient solvers wrapper for the linear solvers --- gtsam/geometry/Cal3DS2.cpp | 30 ++++ gtsam/geometry/Cal3DS2.h | 3 + gtsam/geometry/Cal3_S2.h | 4 +- gtsam/geometry/tests/testCal3DS2.cpp | 10 +- gtsam/nonlinear/GaussNewtonOptimizer.cpp | 20 +-- gtsam/nonlinear/GradientDescentOptimizer.cpp | 138 ------------------ .../nonlinear/LevenbergMarquardtOptimizer.cpp | 35 +---- .../NonlinearConjugateGradientOptimizer.cpp | 64 ++++++++ ... => NonlinearConjugateGradientOptimizer.h} | 114 +++++---------- .../SuccessiveLinearizationOptimizer.cpp | 81 ++++++++++ .../SuccessiveLinearizationOptimizer.h | 42 +----- tests/testGradientDescentOptimizer.cpp | 56 +------ 12 files changed, 236 insertions(+), 361 deletions(-) delete mode 100644 gtsam/nonlinear/GradientDescentOptimizer.cpp create mode 100644 gtsam/nonlinear/NonlinearConjugateGradientOptimizer.cpp rename gtsam/nonlinear/{GradientDescentOptimizer.h => NonlinearConjugateGradientOptimizer.h} (71%) create mode 100644 gtsam/nonlinear/SuccessiveLinearizationOptimizer.cpp diff --git a/gtsam/geometry/Cal3DS2.cpp b/gtsam/geometry/Cal3DS2.cpp index dcb6d6d54..8f15f81d2 100644 --- a/gtsam/geometry/Cal3DS2.cpp +++ b/gtsam/geometry/Cal3DS2.cpp @@ -78,6 +78,36 @@ Point2 Cal3DS2::uncalibrate(const Point2& p, return Point2(fx_* x2 + s_ * y2 + u0_, fy_ * y2 + v0_) ; } +/* ************************************************************************* */ +Point2 Cal3DS2::calibrate(const Point2& pi, const double tol) const { + // Use the following fixed point iteration to invert the radial distortion. + // pn_{t+1} = (inv(K)*pi - dp(pn_{t})) / g(pn_{t}) + + const Point2 invKPi ((1 / fx_) * (pi.x() - u0_ - (s_ / fy_) * (pi.y() - v0_)), + (1 / fy_) * (pi.y() - v0_)); + + // initialize by ignoring the distortion at all, might be problematic for pixels around boundary + Point2 pn = invKPi; + + // iterate until the uncalibrate is close to the actual pixel coordinate + const int maxIterations = 10; + int iteration; + for ( iteration = 0 ; iteration < maxIterations ; ++iteration ) { + if ( uncalibrate(pn).dist(pi) <= tol ) break; + const double x = pn.x(), y = pn.y(), xy = x*y, xx = x*x, yy = y*y ; + const double r = xx + yy ; + const double g = (1+k1_*r+k2_*r*r) ; + const double dx = 2*k3_*xy + k4_*(r+2*xx) ; + const double dy = 2*k4_*xy + k3_*(r+2*yy) ; + pn = (invKPi - Point2(dx,dy))/g ; + } + + if ( iteration >= maxIterations ) + throw std::runtime_error("Cal3DS2::calibrate fails to converge. need a better initialization"); + + return pn; +} + /* ************************************************************************* */ Matrix Cal3DS2::D2d_intrinsic(const Point2& p) const { //const double fx = fx_, fy = fy_, s = s_ ; diff --git a/gtsam/geometry/Cal3DS2.h b/gtsam/geometry/Cal3DS2.h index d26b9a19d..d33b452c1 100644 --- a/gtsam/geometry/Cal3DS2.h +++ b/gtsam/geometry/Cal3DS2.h @@ -96,6 +96,9 @@ public: boost::optional H1 = boost::none, boost::optional H2 = boost::none) const ; + /// Conver a pixel coordinate to ideal coordinate + Point2 calibrate(const Point2& p, const double tol=1e-5) const; + /// Derivative of uncalibrate wrpt intrinsic coordinates Matrix D2d_intrinsic(const Point2& p) const ; diff --git a/gtsam/geometry/Cal3_S2.h b/gtsam/geometry/Cal3_S2.h index 9d6cb4fbf..3c88aaeb2 100644 --- a/gtsam/geometry/Cal3_S2.h +++ b/gtsam/geometry/Cal3_S2.h @@ -136,8 +136,8 @@ namespace gtsam { /// convert image coordinates uv to intrinsic coordinates xy Point2 calibrate(const Point2& p) const { const double u = p.x(), v = p.y(); - return Point2((1 / fx_) * (u - u0_ - (s_ / fy_) * (v - v0_)), (1 / fy_) - * (v - v0_)); + return Point2((1 / fx_) * (u - u0_ - (s_ / fy_) * (v - v0_)), + (1 / fy_) * (v - v0_)); } /// @} diff --git a/gtsam/geometry/tests/testCal3DS2.cpp b/gtsam/geometry/tests/testCal3DS2.cpp index 9f8777e49..c73ae1182 100644 --- a/gtsam/geometry/tests/testCal3DS2.cpp +++ b/gtsam/geometry/tests/testCal3DS2.cpp @@ -29,7 +29,7 @@ static Cal3DS2 K(500, 100, 0.1, 320, 240, 1e-3, 2.0*1e-3, 3.0*1e-3, 4.0*1e-3); static Point2 p(2,3); /* ************************************************************************* */ -TEST( Cal3DS2, calibrate) +TEST( Cal3DS2, uncalibrate) { Vector k = K.k() ; double r = p.x()*p.x() + p.y()*p.y() ; @@ -43,6 +43,14 @@ TEST( Cal3DS2, calibrate) CHECK(assert_equal(q,p_i)); } +TEST( Cal3DS2, calibrate ) +{ + Point2 pn(0.5, 0.5); + Point2 pi = K.uncalibrate(pn); + Point2 pn_hat = K.calibrate(pi); + CHECK( pn.equals(pn_hat, 1e-5)); +} + Point2 uncalibrate_(const Cal3DS2& k, const Point2& pt) { return k.uncalibrate(pt); } /* ************************************************************************* */ diff --git a/gtsam/nonlinear/GaussNewtonOptimizer.cpp b/gtsam/nonlinear/GaussNewtonOptimizer.cpp index c447a39e3..05ba56a21 100644 --- a/gtsam/nonlinear/GaussNewtonOptimizer.cpp +++ b/gtsam/nonlinear/GaussNewtonOptimizer.cpp @@ -16,8 +16,6 @@ * @date Feb 26, 2012 */ -#include -#include #include using namespace std; @@ -32,22 +30,8 @@ void GaussNewtonOptimizer::iterate() { // Linearize graph GaussianFactorGraph::shared_ptr linear = graph_.linearize(current.values, *params_.ordering); - // Optimize - VectorValues delta; - { - if ( params_.isMultifrontal() ) { - delta = GaussianJunctionTree(*linear).optimize(params_.getEliminationFunction()); - } - else if ( params_.isSequential() ) { - delta = gtsam::optimize(*EliminationTree::Create(*linear)->eliminate(params_.getEliminationFunction())); - } - else if ( params_.isCG() ) { - throw runtime_error("todo: "); - } - else { - throw runtime_error("Optimization parameter is invalid: GaussNewtonParams::elimination"); - } - } + // Solve Factor Graph + const VectorValues delta = solveGaussianFactorGraph(*linear, params_); // Maybe show output if(params_.verbosity >= NonlinearOptimizerParams::DELTA) delta.print("delta"); diff --git a/gtsam/nonlinear/GradientDescentOptimizer.cpp b/gtsam/nonlinear/GradientDescentOptimizer.cpp deleted file mode 100644 index 4a471de4d..000000000 --- a/gtsam/nonlinear/GradientDescentOptimizer.cpp +++ /dev/null @@ -1,138 +0,0 @@ -/** - * @file GradientDescentOptimizer.cpp - * @brief - * @author ydjian - * @date Jun 11, 2012 - */ - -#include -#include -#include -#include -#include - -#include - -using namespace std; - -namespace gtsam { - -/** - * Return the gradient vector of a nonlinear factor given a linearization point and a variable ordering - * Can be moved to NonlinearFactorGraph.h if desired - */ -void gradientInPlace(const NonlinearFactorGraph &nfg, const Values &values, const Ordering &ordering, VectorValues &g) { - - // Linearize graph - GaussianFactorGraph::shared_ptr linear = nfg.linearize(values, ordering); - FactorGraph jfg; jfg.reserve(linear->size()); - BOOST_FOREACH(const GaussianFactorGraph::sharedFactor& factor, *linear) { - if(boost::shared_ptr jf = boost::dynamic_pointer_cast(factor)) - jfg.push_back((jf)); - else - jfg.push_back(boost::make_shared(*factor)); - } - - // compute the gradient direction - gradientAtZero(jfg, g); -} - - -/* ************************************************************************* */ -void GradientDescentOptimizer::iterate() { - - - // Pull out parameters we'll use - const NonlinearOptimizerParams::Verbosity nloVerbosity = params_.verbosity; - - // compute the gradient vector - gradientInPlace(graph_, state_.values, *ordering_, *gradient_); - - /* normalize it such that it becomes a unit vector */ - const double g = gradient_->vector().norm(); - gradient_->vector() /= g; - - // perform the golden section search algorithm to decide the the optimal step size - // detail refer to http://en.wikipedia.org/wiki/Golden_section_search - VectorValues step = VectorValues::SameStructure(*gradient_); - const double phi = 0.5*(1.0+std::sqrt(5.0)), resphi = 2.0 - phi, tau = 1e-5; - double minStep = -1.0, maxStep = 0, - newStep = minStep + (maxStep-minStep) / (phi+1.0) ; - - step.vector() = newStep * gradient_->vector(); - Values newValues = state_.values.retract(step, *ordering_); - double newError = graph_.error(newValues); - - if ( nloVerbosity ) { - std::cout << "minStep = " << minStep << ", maxStep = " << maxStep << ", newStep = " << newStep << ", newError = " << newError << std::endl; - } - - while (true) { - const bool flag = (maxStep - newStep > newStep - minStep) ? true : false ; - const double testStep = flag ? - newStep + resphi * (maxStep - newStep) : newStep - resphi * (newStep - minStep); - - if ( (maxStep- minStep) < tau * (std::fabs(testStep) + std::fabs(newStep)) ) { - newStep = 0.5*(minStep+maxStep); - step.vector() = newStep * gradient_->vector(); - newValues = state_.values.retract(step, *ordering_); - newError = graph_.error(newValues); - - if ( newError < state_.error ) { - state_.values = state_.values.retract(step, *ordering_); - state_.error = graph_.error(state_.values); - } - - break; - } - - step.vector() = testStep * gradient_->vector(); - const Values testValues = state_.values.retract(step, *ordering_); - const double testError = graph_.error(testValues); - - // update the working range - if ( testError >= newError ) { - if ( flag ) maxStep = testStep; - else minStep = testStep; - } - else { - if ( flag ) { - minStep = newStep; - newStep = testStep; - newError = testError; - } - else { - maxStep = newStep; - newStep = testStep; - newError = testError; - } - } - - if ( nloVerbosity ) { - std::cout << "minStep = " << minStep << ", maxStep = " << maxStep << ", newStep = " << newStep << ", newError = " << newError << std::endl; - } - } - // Increment the iteration counter - ++state_.iterations; -} - -double ConjugateGradientOptimizer::System::error(const State &state) const { - return graph_.error(state); -} - -ConjugateGradientOptimizer::System::Gradient ConjugateGradientOptimizer::System::gradient(const State &state) const { - Gradient result = state.zeroVectors(ordering_); - gradientInPlace(graph_, state, ordering_, result); - return result; -} -ConjugateGradientOptimizer::System::State ConjugateGradientOptimizer::System::advance(const State ¤t, const double alpha, const Gradient &g) const { - Gradient step = g; - step.vector() *= alpha; - return current.retract(step, ordering_); -} - -Values ConjugateGradientOptimizer::optimize() { - return conjugateGradient(System(graph_, *ordering_), initialEstimate_, params_, !cg_); -} - -} /* namespace gtsam */ diff --git a/gtsam/nonlinear/LevenbergMarquardtOptimizer.cpp b/gtsam/nonlinear/LevenbergMarquardtOptimizer.cpp index 15c2262fa..256763c59 100644 --- a/gtsam/nonlinear/LevenbergMarquardtOptimizer.cpp +++ b/gtsam/nonlinear/LevenbergMarquardtOptimizer.cpp @@ -19,12 +19,9 @@ #include #include +#include #include // For NegativeMatrixException -#include -#include -#include -#include #include #include @@ -106,34 +103,8 @@ void LevenbergMarquardtOptimizer::iterate() { // Try solving try { - - // Optimize - VectorValues delta; - if ( params_.isMultifrontal() ) { - delta = GaussianJunctionTree(dampedSystem).optimize(params_.getEliminationFunction()); - } - else if ( params_.isSequential() ) { - delta = gtsam::optimize(*EliminationTree::Create(dampedSystem)->eliminate(params_.getEliminationFunction())); - } - else if ( params_.isCG() ) { - - if ( !params_.iterativeParams ) throw runtime_error("LMSolver: cg parameter has to be assigned ..."); - - if ( boost::dynamic_pointer_cast(params_.iterativeParams)) { - SimpleSPCGSolver solver (dampedSystem, *boost::dynamic_pointer_cast(params_.iterativeParams)); - delta = solver.optimize(); - } - else if ( boost::dynamic_pointer_cast(params_.iterativeParams) ) { - SubgraphSolver solver (dampedSystem, *boost::dynamic_pointer_cast(params_.iterativeParams)); - delta = solver.optimize(); - } - else { - throw runtime_error("LMSolver: special cg parameter type is not handled in LM solver ..."); - } - } - else { - throw runtime_error("Optimization parameter is invalid: LevenbergMarquardtParams::elimination"); - } + // Solve Damped Gaussian Factor Graph + const VectorValues delta = solveGaussianFactorGraph(dampedSystem, params_); if (lmVerbosity >= LevenbergMarquardtParams::TRYLAMBDA) cout << "linear delta norm = " << delta.vector().norm() << endl; if (lmVerbosity >= LevenbergMarquardtParams::TRYDELTA) delta.print("delta"); diff --git a/gtsam/nonlinear/NonlinearConjugateGradientOptimizer.cpp b/gtsam/nonlinear/NonlinearConjugateGradientOptimizer.cpp new file mode 100644 index 000000000..f0b3852ef --- /dev/null +++ b/gtsam/nonlinear/NonlinearConjugateGradientOptimizer.cpp @@ -0,0 +1,64 @@ +/** + * @file GradientDescentOptimizer.cpp + * @brief + * @author ydjian + * @date Jun 11, 2012 + */ + +#include +#include +#include +#include +#include + +#include + +using namespace std; + +namespace gtsam { + +/* Return the gradient vector of a nonlinear factor given a linearization point and a variable ordering + * Can be moved to NonlinearFactorGraph.h if desired */ +void gradientInPlace(const NonlinearFactorGraph &nfg, const Values &values, const Ordering &ordering, VectorValues &g) { + // Linearize graph + GaussianFactorGraph::shared_ptr linear = nfg.linearize(values, ordering); + FactorGraph jfg; jfg.reserve(linear->size()); + BOOST_FOREACH(const GaussianFactorGraph::sharedFactor& factor, *linear) { + if(boost::shared_ptr jf = boost::dynamic_pointer_cast(factor)) + jfg.push_back((jf)); + else + jfg.push_back(boost::make_shared(*factor)); + } + // compute the gradient direction + gradientAtZero(jfg, g); +} + +double NonlinearConjugateGradientOptimizer::System::error(const State &state) const { + return graph_.error(state); +} + +NonlinearConjugateGradientOptimizer::System::Gradient NonlinearConjugateGradientOptimizer::System::gradient(const State &state) const { + Gradient result = state.zeroVectors(ordering_); + gradientInPlace(graph_, state, ordering_, result); + return result; +} +NonlinearConjugateGradientOptimizer::System::State NonlinearConjugateGradientOptimizer::System::advance(const State ¤t, const double alpha, const Gradient &g) const { + Gradient step = g; + step.vector() *= alpha; + return current.retract(step, ordering_); +} + +void NonlinearConjugateGradientOptimizer::iterate() { + size_t dummy ; + boost::tie(state_.values, dummy) = nonlinearConjugateGradient(System(graph_, *ordering_), state_.values, params_, true /* single iterations */); + ++state_.iterations; + state_.error = graph_.error(state_.values); +} + +const Values& NonlinearConjugateGradientOptimizer::optimize() { + boost::tie(state_.values, state_.iterations) = nonlinearConjugateGradient(System(graph_, *ordering_), state_.values, params_, false /* up to convergent */); + state_.error = graph_.error(state_.values); + return state_.values; +} + +} /* namespace gtsam */ diff --git a/gtsam/nonlinear/GradientDescentOptimizer.h b/gtsam/nonlinear/NonlinearConjugateGradientOptimizer.h similarity index 71% rename from gtsam/nonlinear/GradientDescentOptimizer.h rename to gtsam/nonlinear/NonlinearConjugateGradientOptimizer.h index 2407ee39f..5bac11f5e 100644 --- a/gtsam/nonlinear/GradientDescentOptimizer.h +++ b/gtsam/nonlinear/NonlinearConjugateGradientOptimizer.h @@ -1,7 +1,7 @@ /** * @file GradientDescentOptimizer.cpp * @brief - * @author ydjian + * @author Yong-Dian Jian * @date Jun 11, 2012 */ @@ -9,75 +9,31 @@ #include #include +#include namespace gtsam { -/* an implementation of gradient-descent method using the NLO interface */ - -class GradientDescentState : public NonlinearOptimizerState { - +/** An implementation of the nonlinear cg method using the template below */ +class NonlinearConjugateGradientState : public NonlinearOptimizerState { public: - typedef NonlinearOptimizerState Base; - - GradientDescentState(const NonlinearFactorGraph& graph, const Values& values) + NonlinearConjugateGradientState(const NonlinearFactorGraph& graph, const Values& values) : Base(graph, values) {} }; -class GradientDescentOptimizer : public NonlinearOptimizer { - -public: - - typedef boost::shared_ptr shared_ptr; - typedef NonlinearOptimizer Base; - typedef GradientDescentState States; - typedef NonlinearOptimizerParams Parameters; - -protected: - - Parameters params_; - States state_; - Ordering::shared_ptr ordering_; - VectorValues::shared_ptr gradient_; - -public: - - GradientDescentOptimizer(const NonlinearFactorGraph& graph, const Values& initialValues, const Parameters& params = Parameters()) - : Base(graph), params_(params), state_(graph, initialValues), - ordering_(initialValues.orderingArbitrary()), - gradient_(new VectorValues(initialValues.zeroVectors(*ordering_))) {} - - virtual ~GradientDescentOptimizer() {} - - virtual void iterate(); - -protected: - - virtual const NonlinearOptimizerState& _state() const { return state_; } - virtual const NonlinearOptimizerParams& _params() const { return params_; } -}; - - -/** - * An implementation of the nonlinear cg method using the template below - */ - -class ConjugateGradientOptimizer { - +class NonlinearConjugateGradientOptimizer : public NonlinearOptimizer { + /* a class for the nonlinearConjugateGradient template */ class System { - public: - typedef Values State; typedef VectorValues Gradient; + typedef NonlinearOptimizerParams Parameters; protected: - - NonlinearFactorGraph graph_; - Ordering ordering_; + const NonlinearFactorGraph &graph_; + const Ordering &ordering_; public: - System(const NonlinearFactorGraph &graph, const Ordering &ordering): graph_(graph), ordering_(ordering) {} double error(const State &state) const ; Gradient gradient(const State &state) const ; @@ -85,35 +41,32 @@ class ConjugateGradientOptimizer { }; public: - + typedef NonlinearOptimizer Base; + typedef NonlinearConjugateGradientState States; typedef NonlinearOptimizerParams Parameters; - typedef boost::shared_ptr shared_ptr; + typedef boost::shared_ptr shared_ptr; protected: - - NonlinearFactorGraph graph_; - Values initialEstimate_; + States state_; Parameters params_; Ordering::shared_ptr ordering_; VectorValues::shared_ptr gradient_; - bool cg_; public: - ConjugateGradientOptimizer(const NonlinearFactorGraph& graph, const Values& initialValues, - const Parameters& params = Parameters(), const bool cg = true) - : graph_(graph), initialEstimate_(initialValues), params_(params), - ordering_(initialValues.orderingArbitrary()), - gradient_(new VectorValues(initialValues.zeroVectors(*ordering_))), - cg_(cg) {} + NonlinearConjugateGradientOptimizer(const NonlinearFactorGraph& graph, const Values& initialValues, + const Parameters& params = Parameters()) + : Base(graph), state_(graph, initialValues), params_(params), ordering_(initialValues.orderingArbitrary()), + gradient_(new VectorValues(initialValues.zeroVectors(*ordering_))){} - virtual ~ConjugateGradientOptimizer() {} - virtual Values optimize () ; + virtual ~NonlinearConjugateGradientOptimizer() {} + virtual void iterate(); + virtual const Values& optimize (); + virtual const NonlinearOptimizerState& _state() const { return state_; } + virtual const NonlinearOptimizerParams& _params() const { return params_; } }; -/** - * Implement the golden-section line search algorithm - */ +/** Implement the golden-section line search algorithm */ template double lineSearch(const S &system, const V currentValues, const W &gradient) { @@ -171,18 +124,20 @@ double lineSearch(const S &system, const V currentValues, const W &gradient) { * * The last parameter is a switch between gradient-descent and conjugate gradient */ - template -V conjugateGradient(const S &system, const V &initial, const NonlinearOptimizerParams ¶ms, const bool gradientDescent) { +boost::tuple nonlinearConjugateGradient(const S &system, const V &initial, const NonlinearOptimizerParams ¶ms, const bool singleIteration, const bool gradientDescent = false) { - GTSAM_CONCEPT_MANIFOLD_TYPE(V); + // GTSAM_CONCEPT_MANIFOLD_TYPE(V); + + Index iteration = 0; // check if we're already close enough double currentError = system.error(initial); if(currentError <= params.errorTol) { - if (params.verbosity >= NonlinearOptimizerParams::ERROR) + if (params.verbosity >= NonlinearOptimizerParams::ERROR){ std::cout << "Exiting, as error = " << currentError << " < " << params.errorTol << std::endl; - return initial; + } + return boost::tie(initial, iteration); } V currentValues = initial; @@ -194,14 +149,12 @@ V conjugateGradient(const S &system, const V &initial, const NonlinearOptimizerP double alpha = lineSearch(system, currentValues, direction); currentValues = system.advance(prevValues, alpha, direction); currentError = system.error(currentValues); - Index iteration = 0; // Maybe show output if (params.verbosity >= NonlinearOptimizerParams::ERROR) std::cout << "Initial error: " << currentError << std::endl; // Iterative loop do { - if ( gradientDescent == true) { direction = system.gradient(currentValues); } @@ -222,13 +175,14 @@ V conjugateGradient(const S &system, const V &initial, const NonlinearOptimizerP // Maybe show output if(params.verbosity >= NonlinearOptimizerParams::ERROR) std::cout << "currentError: " << currentError << std::endl; } while( ++iteration < params.maxIterations && + !singleIteration && !checkConvergence(params.relativeErrorTol, params.absoluteErrorTol, params.errorTol, prevError, currentError, params.verbosity)); // Printing if verbose if (params.verbosity >= NonlinearOptimizerParams::ERROR && iteration >= params.maxIterations) - std::cout << "Terminating because reached maximum iterations" << std::endl; + std::cout << "nonlinearConjugateGradient: Terminating because reached maximum iterations" << std::endl; - return currentValues; + return boost::tie(currentValues, iteration); } diff --git a/gtsam/nonlinear/SuccessiveLinearizationOptimizer.cpp b/gtsam/nonlinear/SuccessiveLinearizationOptimizer.cpp new file mode 100644 index 000000000..0fe0440ae --- /dev/null +++ b/gtsam/nonlinear/SuccessiveLinearizationOptimizer.cpp @@ -0,0 +1,81 @@ +/** + * @file SuccessiveLinearizationOptimizer.cpp + * @brief + * @date Jul 24, 2012 + * @author Yong-Dian Jian + */ + +#include +#include +#include +#include +#include +#include +#include +#include + +namespace gtsam { + +void SuccessiveLinearizationParams::print(const std::string& str) const { + NonlinearOptimizerParams::print(str); + switch ( linearSolverType ) { + case MULTIFRONTAL_CHOLESKY: + std::cout << " linear solver type: MULTIFRONTAL CHOLESKY\n"; + break; + case MULTIFRONTAL_QR: + std::cout << " linear solver type: MULTIFRONTAL QR\n"; + break; + case SEQUENTIAL_CHOLESKY: + std::cout << " linear solver type: SEQUENTIAL CHOLESKY\n"; + break; + case SEQUENTIAL_QR: + std::cout << " linear solver type: SEQUENTIAL QR\n"; + break; + case CHOLMOD: + std::cout << " linear solver type: CHOLMOD\n"; + break; + case CG: + std::cout << " linear solver type: CG\n"; + break; + default: + std::cout << " linear solver type: (invalid)\n"; + break; + } + + if(ordering) + std::cout << " ordering: custom\n"; + else + std::cout << " ordering: COLAMD\n"; + + std::cout.flush(); +} + +VectorValues solveGaussianFactorGraph(const GaussianFactorGraph &gfg, const SuccessiveLinearizationParams ¶ms) { + VectorValues delta; + if ( params.isMultifrontal() ) { + delta = GaussianJunctionTree(gfg).optimize(params.getEliminationFunction()); + } + else if ( params.isSequential() ) { + delta = gtsam::optimize(*EliminationTree::Create(gfg)->eliminate(params.getEliminationFunction())); + } + else if ( params.isCG() ) { + if ( !params.iterativeParams ) throw std::runtime_error("solveGaussianFactorGraph: cg parameter has to be assigned ..."); + if ( boost::dynamic_pointer_cast(params.iterativeParams)) { + SimpleSPCGSolver solver (gfg, *boost::dynamic_pointer_cast(params.iterativeParams)); + delta = solver.optimize(); + } + else if ( boost::dynamic_pointer_cast(params.iterativeParams) ) { + SubgraphSolver solver (gfg, *boost::dynamic_pointer_cast(params.iterativeParams)); + delta = solver.optimize(); + } + else { + throw std::runtime_error("solveGaussianFactorGraph: special cg parameter type is not handled in LM solver ..."); + } + } + else { + throw std::runtime_error("solveGaussianFactorGraph: Optimization parameter is invalid"); + } + return delta; +} + +} diff --git a/gtsam/nonlinear/SuccessiveLinearizationOptimizer.h b/gtsam/nonlinear/SuccessiveLinearizationOptimizer.h index 7033613ac..be66f37d4 100644 --- a/gtsam/nonlinear/SuccessiveLinearizationOptimizer.h +++ b/gtsam/nonlinear/SuccessiveLinearizationOptimizer.h @@ -40,43 +40,8 @@ public: IterativeOptimizationParameters::shared_ptr iterativeParams; ///< The container for iterativeOptimization parameters. used in CG Solvers. SuccessiveLinearizationParams() : linearSolverType(MULTIFRONTAL_CHOLESKY) {} - virtual ~SuccessiveLinearizationParams() {} - virtual void print(const std::string& str = "") const { - NonlinearOptimizerParams::print(str); - switch ( linearSolverType ) { - case MULTIFRONTAL_CHOLESKY: - std::cout << " linear solver type: MULTIFRONTAL CHOLESKY\n"; - break; - case MULTIFRONTAL_QR: - std::cout << " linear solver type: MULTIFRONTAL QR\n"; - break; - case SEQUENTIAL_CHOLESKY: - std::cout << " linear solver type: SEQUENTIAL CHOLESKY\n"; - break; - case SEQUENTIAL_QR: - std::cout << " linear solver type: SEQUENTIAL QR\n"; - break; - case CHOLMOD: - std::cout << " linear solver type: CHOLMOD\n"; - break; - case CG: - std::cout << " linear solver type: CG\n"; - break; - default: - std::cout << " linear solver type: (invalid)\n"; - break; - } - - if(ordering) - std::cout << " ordering: custom\n"; - else - std::cout << " ordering: COLAMD\n"; - - std::cout.flush(); - } - inline bool isMultifrontal() const { return (linearSolverType == MULTIFRONTAL_CHOLESKY) || (linearSolverType == MULTIFRONTAL_QR); } @@ -93,7 +58,9 @@ public: return (linearSolverType == CG); } - GaussianFactorGraph::Eliminate getEliminationFunction() { + virtual void print(const std::string& str) const; + + GaussianFactorGraph::Eliminate getEliminationFunction() const { switch (linearSolverType) { case MULTIFRONTAL_CHOLESKY: case SEQUENTIAL_CHOLESKY: @@ -111,4 +78,7 @@ public: } }; +/* a wrapper for solving a GaussianFactorGraph according to the parameters */ +VectorValues solveGaussianFactorGraph(const GaussianFactorGraph &gfg, const SuccessiveLinearizationParams ¶ms) ; + } /* namespace gtsam */ diff --git a/tests/testGradientDescentOptimizer.cpp b/tests/testGradientDescentOptimizer.cpp index fefa2a05f..bb7743bd9 100644 --- a/tests/testGradientDescentOptimizer.cpp +++ b/tests/testGradientDescentOptimizer.cpp @@ -6,7 +6,7 @@ */ #include -#include +#include #include @@ -51,31 +51,6 @@ boost::tuple generateProblem() { return boost::tie(graph, initialEstimate); } - -/* ************************************************************************* */ -TEST(optimize, GradientDescentOptimizer) { - - pose2SLAM::Graph graph ; - pose2SLAM::Values initialEstimate; - - boost::tie(graph, initialEstimate) = generateProblem(); - // cout << "initial error = " << graph.error(initialEstimate) << endl ; - - // Single Step Optimization using Levenberg-Marquardt - NonlinearOptimizerParams param; - param.maxIterations = 500; /* requires a larger number of iterations to converge */ - param.verbosity = NonlinearOptimizerParams::SILENT; - - GradientDescentOptimizer optimizer(graph, initialEstimate, param); - Values result = optimizer.optimize(); -// cout << "gd1 solver final error = " << graph.error(result) << endl; - - /* the optimality of the solution is not comparable to the */ - DOUBLES_EQUAL(0.0, graph.error(result), 1e-2); - - CHECK(1); -} - /* ************************************************************************* */ TEST(optimize, ConjugateGradientOptimizer) { @@ -90,8 +65,7 @@ TEST(optimize, ConjugateGradientOptimizer) { param.maxIterations = 500; /* requires a larger number of iterations to converge */ param.verbosity = NonlinearOptimizerParams::SILENT; - - ConjugateGradientOptimizer optimizer(graph, initialEstimate, param, true); + NonlinearConjugateGradientOptimizer optimizer(graph, initialEstimate, param); Values result = optimizer.optimize(); // cout << "cg final error = " << graph.error(result) << endl; @@ -99,32 +73,6 @@ TEST(optimize, ConjugateGradientOptimizer) { DOUBLES_EQUAL(0.0, graph.error(result), 1e-2); } -/* ************************************************************************* */ -TEST(optimize, GradientDescentOptimizer2) { - - pose2SLAM::Graph graph ; - pose2SLAM::Values initialEstimate; - - boost::tie(graph, initialEstimate) = generateProblem(); -// cout << "initial error = " << graph.error(initialEstimate) << endl ; - - // Single Step Optimization using Levenberg-Marquardt - NonlinearOptimizerParams param; - param.maxIterations = 500; /* requires a larger number of iterations to converge */ - param.verbosity = NonlinearOptimizerParams::SILENT; - - - ConjugateGradientOptimizer optimizer(graph, initialEstimate, param, false); - Values result = optimizer.optimize(); -// cout << "gd2 solver final error = " << graph.error(result) << endl; - - /* the optimality of the solution is not comparable to the */ - DOUBLES_EQUAL(0.0, graph.error(result), 1e-2); -} - - - - /* ************************************************************************* */ int main() { TestResult tr; return TestRegistry::runAllTests(tr); } /* ************************************************************************* */