| 1 | /* ---------------------------------------------------------------------------- |
| 2 | |
| 3 | * GTSAM Copyright 2010, Georgia Tech Research Corporation, |
| 4 | * Atlanta, Georgia 30332-0415 |
| 5 | * All Rights Reserved |
| 6 | * Authors: Frank Dellaert, et al. (see THANKS for the full author list) |
| 7 | |
| 8 | * See LICENSE for the license information |
| 9 | |
| 10 | * -------------------------------------------------------------------------- */ |
| 11 | |
| 12 | /** |
| 13 | * @file testGaussMarkov1stOrderFactor.cpp |
| 14 | * @brief Unit tests for the GaussMarkov1stOrder factor |
| 15 | * @author Vadim Indelman |
| 16 | * @date Jan 17, 2012 |
| 17 | */ |
| 18 | |
| 19 | #include <CppUnitLite/TestHarness.h> |
| 20 | #include <gtsam/base/Vector.h> |
| 21 | #include <gtsam/base/numericalDerivative.h> |
| 22 | #include <gtsam/inference/Key.h> |
| 23 | #include <gtsam/nonlinear/Values.h> |
| 24 | #include <gtsam_unstable/slam/GaussMarkov1stOrderFactor.h> |
| 25 | |
| 26 | using namespace std::placeholders; |
| 27 | using namespace std; |
| 28 | using namespace gtsam; |
| 29 | |
| 30 | //! Factors |
| 31 | typedef GaussMarkov1stOrderFactor<Vector3> GaussMarkovFactor; |
| 32 | |
| 33 | /* ************************************************************************* */ |
| 34 | Vector predictionError(const Vector& v1, const Vector& v2, |
| 35 | const GaussMarkovFactor factor) { |
| 36 | return factor.evaluateError(x: v1, x: v2); |
| 37 | } |
| 38 | |
| 39 | /* ************************************************************************* */ |
| 40 | TEST( GaussMarkovFactor, equals ) |
| 41 | { |
| 42 | // Create two identical factors and make sure they're equal |
| 43 | Key x1(1); |
| 44 | Key x2(2); |
| 45 | double delta_t = 0.10; |
| 46 | Vector tau = Vector3(100.0, 150.0, 10.0); |
| 47 | SharedGaussian model = noiseModel::Isotropic::Sigma(dim: 3, sigma: 1.0); |
| 48 | |
| 49 | GaussMarkovFactor factor1(x1, x2, delta_t, tau, model); |
| 50 | GaussMarkovFactor factor2(x1, x2, delta_t, tau, model); |
| 51 | |
| 52 | CHECK(assert_equal(factor1, factor2)); |
| 53 | } |
| 54 | |
| 55 | /* ************************************************************************* */ |
| 56 | TEST( GaussMarkovFactor, error ) |
| 57 | { |
| 58 | Values linPoint; |
| 59 | Key x1(1); |
| 60 | Key x2(2); |
| 61 | double delta_t = 0.10; |
| 62 | Vector3 tau(100.0, 150.0, 10.0); |
| 63 | SharedGaussian model = noiseModel::Isotropic::Sigma(dim: 3, sigma: 1.0); |
| 64 | |
| 65 | Vector3 v1(10.0, 12.0, 13.0); |
| 66 | Vector3 v2(10.0, 15.0, 14.0); |
| 67 | |
| 68 | // Create two nodes |
| 69 | linPoint.insert(j: x1, val: v1); |
| 70 | linPoint.insert(j: x2, val: v2); |
| 71 | |
| 72 | GaussMarkovFactor factor(x1, x2, delta_t, tau, model); |
| 73 | Vector3 error1 = factor.evaluateError(x: v1, x: v2); |
| 74 | |
| 75 | // Manually calculate the error |
| 76 | Vector3 alpha(tau.size()); |
| 77 | Vector3 alpha_v1(tau.size()); |
| 78 | for(int i=0; i<tau.size(); i++){ |
| 79 | alpha(i) = exp(x: - 1/tau(i)*delta_t ); |
| 80 | alpha_v1(i) = alpha(i) * v1(i); |
| 81 | } |
| 82 | Vector3 error2 = v2 - alpha_v1; |
| 83 | |
| 84 | CHECK(assert_equal(error1, error2, 1e-8)); |
| 85 | } |
| 86 | |
| 87 | /* ************************************************************************* */ |
| 88 | TEST (GaussMarkovFactor, jacobian ) { |
| 89 | |
| 90 | Values linPoint; |
| 91 | Key x1(1); |
| 92 | Key x2(2); |
| 93 | double delta_t = 0.10; |
| 94 | Vector3 tau(100.0, 150.0, 10.0); |
| 95 | SharedGaussian model = noiseModel::Isotropic::Sigma(dim: 3, sigma: 1.0); |
| 96 | |
| 97 | GaussMarkovFactor factor(x1, x2, delta_t, tau, model); |
| 98 | |
| 99 | // Update the linearization point |
| 100 | Vector3 v1_upd(0.5, -0.7, 0.3); |
| 101 | Vector3 v2_upd(-0.7, 0.4, 0.9); |
| 102 | |
| 103 | // Calculate the Jacobian matrix using the factor |
| 104 | Matrix computed_H1, computed_H2; |
| 105 | factor.evaluateError(x: v1_upd, x: v2_upd, H&: computed_H1, H&: computed_H2); |
| 106 | |
| 107 | // Calculate the Jacobian matrices H1 and H2 using the numerical derivative function |
| 108 | Matrix numerical_H1, numerical_H2; |
| 109 | numerical_H1 = numericalDerivative21<Vector3, Vector3, Vector3>( |
| 110 | h: std::bind(f: &predictionError, args: std::placeholders::_1, args: std::placeholders::_2, |
| 111 | args&: factor), |
| 112 | x1: v1_upd, x2: v2_upd); |
| 113 | numerical_H2 = numericalDerivative22<Vector3, Vector3, Vector3>( |
| 114 | h: std::bind(f: &predictionError, args: std::placeholders::_1, args: std::placeholders::_2, |
| 115 | args&: factor), |
| 116 | x1: v1_upd, x2: v2_upd); |
| 117 | |
| 118 | // Verify they are equal for this choice of state |
| 119 | CHECK(assert_equal(numerical_H1, computed_H1, 1e-9)); |
| 120 | CHECK(assert_equal(numerical_H2, computed_H2, 1e-9)); |
| 121 | } |
| 122 | |
| 123 | /* ************************************************************************* */ |
| 124 | int main() |
| 125 | { |
| 126 | TestResult tr; return TestRegistry::runAllTests(result&: tr); |
| 127 | } |
| 128 | /* ************************************************************************* */ |
| 129 | |
| 130 | |