1// Ceres Solver - A fast non-linear least squares minimizer
2// Copyright 2010, 2011, 2012 Google Inc. All rights reserved.
3// http://code.google.com/p/ceres-solver/
4//
5// Redistribution and use in source and binary forms, with or without
6// modification, are permitted provided that the following conditions are met:
7//
8// * Redistributions of source code must retain the above copyright notice,
9//   this list of conditions and the following disclaimer.
10// * Redistributions in binary form must reproduce the above copyright notice,
11//   this list of conditions and the following disclaimer in the documentation
12//   and/or other materials provided with the distribution.
13// * Neither the name of Google Inc. nor the names of its contributors may be
14//   used to endorse or promote products derived from this software without
15//   specific prior written permission.
16//
17// THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
18// AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
19// IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
20// ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
21// LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
22// CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
23// SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
24// INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
25// CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
26// ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
27// POSSIBILITY OF SUCH DAMAGE.
28//
29// Author: wjr@google.com (William Rucklidge)
30//
31// Tests for the conditioned cost function.
32
33#include "ceres/conditioned_cost_function.h"
34
35#include "ceres/internal/eigen.h"
36#include "ceres/normal_prior.h"
37#include "ceres/types.h"
38#include "gtest/gtest.h"
39
40namespace ceres {
41namespace internal {
42
43// The size of the cost functions we build.
44static const int kTestCostFunctionSize = 3;
45
46// A simple cost function: return ax + b.
47class LinearCostFunction : public CostFunction {
48 public:
49  LinearCostFunction(double a, double b) : a_(a), b_(b) {
50    set_num_residuals(1);
51    mutable_parameter_block_sizes()->push_back(1);
52  }
53
54  virtual bool Evaluate(double const* const* parameters,
55                        double* residuals,
56                        double** jacobians) const {
57    *residuals = **parameters * a_ + b_;
58    if (jacobians && *jacobians) {
59      **jacobians = a_;
60    }
61
62    return true;
63  }
64
65 private:
66  const double a_, b_;
67};
68
69// Tests that ConditionedCostFunction does what it's supposed to.
70TEST(CostFunctionTest, ConditionedCostFunction) {
71  double v1[kTestCostFunctionSize], v2[kTestCostFunctionSize],
72      jac[kTestCostFunctionSize * kTestCostFunctionSize],
73      result[kTestCostFunctionSize];
74
75  for (int i = 0; i < kTestCostFunctionSize; i++) {
76    v1[i] = i;
77    v2[i] = i * 10;
78    // Seed a few garbage values in the Jacobian matrix, to make sure that
79    // they're overwritten.
80    jac[i * 2] = i * i;
81    result[i] = i * i * i;
82  }
83
84  // Make a cost function that computes x - v2
85  VectorRef v2_vector(v2, kTestCostFunctionSize, 1);
86  Matrix identity(kTestCostFunctionSize, kTestCostFunctionSize);
87  identity.setIdentity();
88  NormalPrior* difference_cost_function = new NormalPrior(identity, v2_vector);
89
90  vector<CostFunction*> conditioners;
91  for (int i = 0; i < kTestCostFunctionSize; i++) {
92    conditioners.push_back(new LinearCostFunction(i + 2, i * 7));
93  }
94
95  ConditionedCostFunction conditioned_cost_function(difference_cost_function,
96                                                    conditioners,
97                                                    TAKE_OWNERSHIP);
98  EXPECT_EQ(difference_cost_function->num_residuals(),
99            conditioned_cost_function.num_residuals());
100  EXPECT_EQ(difference_cost_function->parameter_block_sizes(),
101            conditioned_cost_function.parameter_block_sizes());
102
103  double *parameters[1];
104  parameters[0] = v1;
105  double *jacs[1];
106  jacs[0] = jac;
107
108  conditioned_cost_function.Evaluate(parameters, result, jacs);
109  for (int i = 0; i < kTestCostFunctionSize; i++) {
110    EXPECT_DOUBLE_EQ((i + 2) * (v1[i] - v2[i]) + i * 7, result[i]);
111  }
112
113  for (int i = 0; i < kTestCostFunctionSize; i++) {
114    for (int j = 0; j < kTestCostFunctionSize; j++) {
115      double actual = jac[i * kTestCostFunctionSize + j];
116      if (i != j) {
117        EXPECT_DOUBLE_EQ(0, actual);
118      } else {
119        EXPECT_DOUBLE_EQ(i + 2, actual);
120      }
121    }
122  }
123}
124
125}  // namespace internal
126}  // namespace ceres
127