ScenarioRunner.cpp
Go to the documentation of this file.
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 
18 #include <gtsam/base/timing.h>
20 
21 #include <cmath>
22 
23 using namespace std;
24 
25 namespace gtsam {
26 
27 static double intNoiseVar = 0.0000001;
28 static const Matrix3 kIntegrationErrorCovariance = intNoiseVar * I_3x3;
29 
30 PreintegratedImuMeasurements ScenarioRunner::integrate(
31  double T, const Bias& estimatedBias, bool corrupted) const {
32  gttic_(integrate);
33  PreintegratedImuMeasurements pim(p_, estimatedBias);
34 
35  const double dt = imuSampleTime();
36  const size_t nrSteps = T / dt;
37  double t = 0;
38  for (size_t k = 0; k < nrSteps; k++, t += dt) {
40  corrupted ? measuredAngularVelocity(t) : actualAngularVelocity(t);
42  corrupted ? measuredSpecificForce(t) : actualSpecificForce(t);
44  }
45 
46  return pim;
47 }
48 
49 NavState ScenarioRunner::predict(const PreintegratedImuMeasurements& pim,
50  const Bias& estimatedBias) const {
51  const NavState state_i(scenario_.pose(0), scenario_.velocity_n(0));
52  return pim.predict(state_i, estimatedBias);
53 }
54 
55 Matrix9 ScenarioRunner::estimateCovariance(double T, size_t N,
56  const Bias& estimatedBias) const {
57  gttic_(estimateCovariance);
58 
59  // Get predict prediction from ground truth measurements
60  NavState prediction = predict(integrate(T));
61 
62  // Sample !
63  Matrix samples(9, N);
64  Vector9 sum = Vector9::Zero();
65  for (size_t i = 0; i < N; i++) {
66  auto pim = integrate(T, estimatedBias, true);
67  NavState sampled = predict(pim);
68  Vector9 xi = sampled.localCoordinates(prediction);
69  samples.col(i) = xi;
70  sum += xi;
71  }
72 
73  // Compute MC covariance
74  Vector9 sampleMean = sum / N;
75  Matrix9 Q;
76  Q.setZero();
77  for (size_t i = 0; i < N; i++) {
78  Vector9 xi = samples.col(i) - sampleMean;
79  Q += xi * xi.transpose();
80  }
81 
82  return Q / (N - 1);
83 }
84 
85 Matrix6 ScenarioRunner::estimateNoiseCovariance(size_t N) const {
86  Matrix samples(6, N);
87  Vector6 sum = Vector6::Zero();
88  for (size_t i = 0; i < N; i++) {
89  samples.col(i) << accSampler_.sample() / sqrt_dt_,
90  gyroSampler_.sample() / sqrt_dt_;
91  sum += samples.col(i);
92  }
93 
94  // Compute MC covariance
95  Vector6 sampleMean = sum / N;
96  Matrix6 Q;
97  Q.setZero();
98  for (size_t i = 0; i < N; i++) {
99  Vector6 xi = samples.col(i) - sampleMean;
100  Q += xi * xi.transpose();
101  }
102 
103  return Q / (N - 1);
104 }
105 
106 PreintegratedCombinedMeasurements CombinedScenarioRunner::integrate(
107  double T, const Bias& estimatedBias, bool corrupted) const {
108  gttic_(integrate);
109  PreintegratedCombinedMeasurements pim(p_, estimatedBias, preintMeasCov_);
110 
111  const double dt = imuSampleTime();
112  const size_t nrSteps = T / dt;
113  double t = 0;
114  for (size_t k = 0; k < nrSteps; k++, t += dt) {
116  corrupted ? measuredAngularVelocity(t) : actualAngularVelocity(t);
118  corrupted ? measuredSpecificForce(t) : actualSpecificForce(t);
120  }
121 
122  return pim;
123 }
124 
125 NavState CombinedScenarioRunner::predict(
127  const Bias& estimatedBias) const {
128  const NavState state_i(scenario().pose(0), scenario().velocity_n(0));
129  return pim.predict(state_i, estimatedBias);
130 }
131 
132 Eigen::Matrix<double, 15, 15> CombinedScenarioRunner::estimateCovariance(
133  double T, size_t N, const Bias& estimatedBias) const {
134  gttic_(estimateCovariance);
135 
136  // Get predict prediction from ground truth measurements
137  NavState prediction = predict(integrate(T));
138 
139  // Sample !
140  Matrix samples(15, N);
141  Vector15 sum = Vector15::Zero();
142  for (size_t i = 0; i < N; i++) {
143  auto pim = integrate(T, estimatedBias, true);
144  NavState sampled = predict(pim);
145  Vector15 xi = Vector15::Zero();
146  xi << sampled.localCoordinates(prediction),
147  (estimatedBias_.vector() - estimatedBias.vector());
148  samples.col(i) = xi;
149  sum += xi;
150  }
151 
152  // Compute MC covariance
153  Vector15 sampleMean = sum / N;
155  Q.setZero();
156  for (size_t i = 0; i < N; i++) {
157  Vector15 xi = samples.col(i) - sampleMean;
158  Q += xi * xi.transpose();
159  }
160 
161  return Q / (N - 1);
162 }
163 
164 } // namespace gtsam
timing.h
Timing utilities.
pose
static const Pose3 pose(Rot3(Vector3(1, -1, -1).asDiagonal()), Point3(0, 0, 0.5))
ScenarioRunner.h
Simple class to test navigation scenarios.
screwPose2::xi
Vector xi
Definition: testPose2.cpp:148
gtsam::PreintegratedImuMeasurements::integrateMeasurement
void integrateMeasurement(const Vector3 &measuredAcc, const Vector3 &measuredOmega, const double dt) override
Definition: ImuFactor.cpp:53
dt
const double dt
Definition: testVelocityConstraint.cpp:15
gtsam::NavState
Definition: NavState.h:34
gtsam::Matrix
Eigen::MatrixXd Matrix
Definition: base/Matrix.h:39
gtsam::Vector3
Eigen::Vector3d Vector3
Definition: Vector.h:43
gtsam::imuBias::ConstantBias::vector
Vector6 vector() const
Definition: ImuBias.h:59
Q
Quaternion Q
Definition: testQuaternion.cpp:27
gtsam::PreintegratedCombinedMeasurements::integrateMeasurement
void integrateMeasurement(const Vector3 &measuredAcc, const Vector3 &measuredOmega, const double dt) override
Definition: CombinedImuFactor.cpp:106
samples
set noclip points set clip one set noclip two set bar set border lt lw set xdata set ydata set zdata set x2data set y2data set boxwidth set dummy y set format x g set format y g set format x2 g set format y2 g set format z g set angles radians set nogrid set key title set key left top Right noreverse box linetype linewidth samplen spacing width set nolabel set noarrow set nologscale set logscale x set set pointsize set encoding default set nopolar set noparametric set set samples
Definition: gnuplot_common_settings.hh:32
biased_x_rotation::measuredOmega
const Vector3 measuredOmega
Definition: testPreintegratedRotation.cpp:35
gttic_
#define gttic_(label)
Definition: timing.h:245
gtsam::PreintegratedImuMeasurements
Definition: ImuFactor.h:68
gtsam::kIntegrationErrorCovariance
static const Matrix3 kIntegrationErrorCovariance
Definition: ScenarioRunner.cpp:28
gtsam::PreintegrationBase::predict
NavState predict(const NavState &state_i, const imuBias::ConstantBias &bias_i, OptionalJacobian< 9, 9 > H1={}, OptionalJacobian< 9, 6 > H2={}) const
Predict state at time j.
Definition: PreintegrationBase.cpp:115
forward::scenario
ConstantTwistScenario scenario(Z_3x1, Vector3(v, 0, 0))
Eigen::Triplet< double >
gtsam::imuBias::ConstantBias
Definition: ImuBias.h:32
Eigen::Quaternion
The quaternion class used to represent 3D orientations and rotations.
Definition: ForwardDeclarations.h:293
gtsam
traits
Definition: SFMdata.h:40
gtsam::intNoiseVar
static double intNoiseVar
Definition: ScenarioRunner.cpp:27
std
Definition: BFloat16.h:88
common::measuredAcc
static const Vector3 measuredAcc
Definition: testImuFactor.cpp:181
Eigen::Matrix< double, 9, 9 >
N
#define N
Definition: igam.h:9
gtsam::NavState::localCoordinates
Vector9 localCoordinates(const NavState &g, OptionalJacobian< 9, 9 > H1={}, OptionalJacobian< 9, 9 > H2={}) const
localCoordinates with optional derivatives
Definition: NavState.cpp:134
gtsam::PreintegratedCombinedMeasurements
Definition: CombinedImuFactor.h:66
align_3::t
Point2 t(10, 10)
i
int i
Definition: BiCGSTAB_step_by_step.cpp:9


gtsam
Author(s):
autogenerated on Sat Nov 16 2024 04:04:02